82649 SOUTH EAST EUROPE MUNICIPAL FINANCE REVIEW Local Government Finance in the Western Balkans SEPTEMBER 2013 EUROPE AND CENTRAL ASIA Report No.: 82649-ECA South East Europe Municipal Finance Review Local Government Finance in the Western Balkans September 2013 WORLD BANK, SUSTAINABLE DEVELOPMENT DEPARTMENT, EUROPE AND CENTRAL ASIA REGION (ECSSD) Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete infor- mation to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. ii  Contents Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Acronyms and Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Executive Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi 1. Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Background.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Objective and methodology.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2. Overview: Decentralization in the Western Balkans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1 Administrative structure and functions of sub-national governments.. . . . . . . . . . . . . . . . . . . . 5 2.2 Size and structure of municipalities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3. Local Government Finance: Structure and Composition of Sub-National Budgets.. . . . . . . . . . . . 15 3.1 Weight and institutional structure of local government finance.. . . . . . . . . . . . . . . . . . . . . . . . 15 3.1.1 Importance of local governments in the SEE region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.1.2 Institutional arrangements of local government revenues.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2 Local government revenues.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.3 Local government expenditure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.4 Grants, transfers and inter-governmental fiscal relations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.5 Sub-national borrowing and debt. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4. Key Trends and Findings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.1 Variation and disparities across types of municipalities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.1.1 Population size.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.1.2 Urban and rural.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 4.1.3 Densities.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 4.2 Concentration in capital investment expenditure.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.3 The role of capital cities.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 4.4 Local government revenue and expenditure trends during the period 2008–2011. . . . . . . . . 46 4.5 Inequality, equalization, and selected trends by types of municipalities. ................. 49 4.5.1 Trends of increasing disparities and convergence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4.5.2 Equalization of expenditure needs and fiscal capacity.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4.5.3 Selected trends: who benefitted, who fell behind? .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 5. Conclusion and Directons for a Possible Way Forward.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 7 6. References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1 Annexes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Annex 1: Summary Overview Graphs.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Annex 2: Data Gap Analysis and Reclassification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Annex 3: Colleciton of Local Government Finance Data.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 iii South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLES Table 1: Overview of Sub-National Government Functional Assignments in South-East Europe.. . . . . . 9 Table 2: Local Government Units in SEE countries and population distribution.. . . . . . . . . . . . . . . . . . . 12 Table 3: Mean density (inhabitants per km2) by size of municipalities. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Table 4: Main features of local government finance in 2011 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Table 5: Overview of Sub-National Government Revenue Assignments in South-East Europe .. . . . . . 20 Table 6: Trends in total revenue, own source revenue and grants and transfers.. . . . . . . . . . . . . . . . . . . . 23 Table 7: Trends in own source revenue for the 2008–2011* period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Table 8: Regional comparison of local expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Table 9: Real Growth Rate Index of sources of local revenues – Regional comparison.. . . . . . . . . . . . . . 28 Table 10: Arrangements on Grants and Transfers in SEE.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Table 11: Sub-national borrowing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Table 12: Local per capita expenditures – 2011.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Table 13: Per capita general grants by size of local government unit 2011 (FBiH: 2010). . . . . . . . . . . . . . . 54 Table 14: Per capita expenditure, own source, grants and transfers and total revenue by size of local government unit 2011. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 FIGURES Figure 1: Distribution of municipalities across different sizes (left) and population distribution across different size municipalities (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Figure 2: Composition and evolution of local government revenues.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv Figure 3: Geographic concentration of total revenues by city size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Figure 4: Variations on per capita expenditures in General Public Services within municipality ranges – Kosovo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Figure 5: Concentration of expenditures and population in the capital city.. . . . . . . . . . . . . . . . . . . . . . . xix Figure 6: Per-capita expenditure trends in poorest (1st Quintile) and richest (5th Quintile) municipalities for Kosovo (left) and Montenegro (right) (reference year: 2008).. . . . . . . . . . . . . xx Figure 7: Distribution of the population across different municipalities sizes. . . . . . . . . . . . . . . . . . . . . . . 11 Figure 8: Distribution of municipalities across different sizes.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 9: Grants and transfers still represent a considerable share of sub-national revenues.. . . . . . . . . . 19 Figure 10: Trends in total local revenues – regional comparison.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 11: Trends in share of total revenue from own source revenues – regional comparison.. . . . . . . . . 23 Figure 12: Trends in own source revenues – regional comparison. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Figure 13: Composition of sub-national own source revenues in the SEE Region.. . . . . . . . . . . . . . . . . . . . 24 Figure 14: Trends in total revenue as share of GDP – regional comparison.. . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 15: Trends in total local expenditures as share of GDP – regional comparison.. . . . . . . . . . . . . . . . . 25 Figure 16: Composition of sub-national expenditure by function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Figure 17: Real Growth Rate Index of Grants and Transfers – Regional comparison.. . . . . . . . . . . . . . . . . . 27 Figure 18: Variations in local revenues by population size of municipalities (Macedonia and Montenegro). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Figure 19: Variations in own source revenue (left) and mean per capita tax revenue (right) by city size in Albania. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 20: Geographic concentration of total revenues by city size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 21: Composition of own-source revenues by size of municipalities (Macedonia and Montenegro). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 iv  Figure 22: Disparities of functional expenditure by size of municipalities in Albania and Kosovo.. . . . . . 37 Figure 23: Variations in education expenditures by city size in Croatia.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Figure 24: Variations in housing expenditures by city size in Macedonia.. . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Figure 25: Variations on per capita expenditures in General Public Services within municipality ranges – Kosovo.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Figure 26: Regional variation on per capita own source revenue, total revenue, grants and transfers and capital expenditures.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Figure 27: Variations between urban and rural municipalities in Macedonia.. . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 28: Density versus Environmental Protection per capita expenditures in Macedonia – 2011.. . . . . 41 Figure 29: Density versus Education per capita expenditures in Macedonia – 2011. . . . . . . . . . . . . . . . . . . 41 Figure 30: Regional variations on capital expenditures by city size and total share.. . . . . . . . . . . . . . . . . . . 42 Figure 31: Macedonia: variations on capital expenditures, Skopje dominating .. . . . . . . . . . . . . . . . . . . . . . 43 Figure 32: Albania and Kosovo: variations on capital expenditures, three different stories.. . . . . . . . . . . . 44 Figure 33: Concentration of expenditures and population in the capital city.. . . . . . . . . . . . . . . . . . . . . . . . 45 Figure 34: Concentration of expenditures for different services in capital cities.. . . . . . . . . . . . . . . . . . . . . . 47 Figure 35: Regional trends in local Governments expenditures 2008–2011.. . . . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 36: Regional trends in local Governments capital expenditures 2008–2011.. . . . . . . . . . . . . . . . . . . . 48 Figure 37: Local revenue trends in Montenegro 2008–2011. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 38: Converging trend in per-capita expenditures for poorest (1st Quintile) and richest (5th Quintile) – 2008 reference year.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 39: Diverging trends in per-capita expenditures for poorest (1st Quintile) and richest (5th Quintile) – 2008 reference year.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 40: Trends in functional expenditures for poorest (1st Quintile) and richest (5th Quintile) in Kosovo – 2008 reference year. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 41: Trends in functional expenditures for poorest (1st Quintile) and richest (5th Quintile) in Albania – 2008 reference year.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Figure 42: Regional trends of Local Governments’ Gini coefficient – per capita Total Revenue. ......... 52 Figure 43: Regional trends of Local Governments’ Gini coefficient – per capita Own Source Revenue.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Figure 44: Regional trends of Local Governments’ Gini coefficient – Own Source Revenue. . . . . . . . . . . . 53 Figure 45: Regional trends of Local Governments’ Gini coefficient – Total Revenue.. . . . . . . . . . . . . . . . . . 53 Figure 46: Changes in total revenues by size of municipalities (Albania and Montenegro).. . . . . . . . . . . . 55 Figure 47: Changes in total revenues, own source, and grants and transfers (urban vs. rural) in Macedonia.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Figure 48: Revenue and expenditure trends by size of municipalities (Macedonia). . . . . . . . . . . . . . . . . . . 56 Figure 49: Visual presentation of municipal finance indicators in Croatia.. . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Figure 50: Variations in total revenue sources by city size (As percentage of total).. . . . . . . . . . . . . . . . . . . 66 Figure 51: Variations in total revenue sources by city size (Absolute values).. . . . . . . . . . . . . . . . . . . . . . . . 68 Figure 52: Variations in own source revenue sources by city size (as percentage of own source).. . . . . . . 70 Figure 53: Variations in own source revenue sources by city size (Absolute values).. . . . . . . . . . . . . . . . . . 71 Figure 54: Functional expenditures by city size (As percentage of total expenditures).. . . . . . . . . . . . . . . . 72 Figure 55: Functional expenditures by city size (Absolute values).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Figure 56: Per capita expenditure in Environmental Protection, Education and Housing against population density – Croatia, Kosovo and Macedonia.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Figure 57: Diverging and Converging trends of per capita expenditures between “rich” and “poor municipalities”.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 v South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans BOXES Box 1: Classification of local government revenue sources.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Box 2: The implementation and application of the MFR in the Federation of BiH.. . . . . . . . . . . . . . . . . . . 59 vi Acknowledgements This report was prepared as one of the analytical different stages of preparation and implementa- activities embedded in a regional capacity building tion, in particular Stephen Karam, Catherine program aimed at strengthening the capacity of Farvacque-Vitkovic, Bekim Ymeri, and Maha local governments and local government associa- Armaly. The collaboration from Sabine Palmreuther tions in the South East Europe (SEE) Region: the is gratefully acknowledged. Peer reviewers were World Bank-Austria Urban Partnership Program Ahmed Eiweida, Mihaly Kopoyani, William (UPP). UPP is being implemented in close collabora- Dillinger, Sebastian Eckardt, and Xiaolan Wang. tion between the World Bank’s Europe and Central Asia (ECA) Region Sustainable Cities Initiative The team is indebted to officials and representa- (SCI) at the ECA Urban, Water and Sanitation, and tives from central and local government agencies Disaster Risk Management Unit (ECSUW), and the and local government associations in Albania, World Bank Institute’s (WBI) Urban Practice with Bosnia and Herzegovina, Croatia, Kosovo, FYR funding provided by the Austria Ministry of Finance. Macedonia, Montenegro, and Serbia for providing key inputs, data and help with the analysis. The This study was developed under the guidance team is particularly grateful for the discussions and of Jane Armitage and Ellen Goldstein (Country close collaboration with the Network of Associations Directors), and prepared under the direction of Wael of Local Authorities of South-East Europe (NALAS), Zakout and Sumila Gulyani (Sector Managers). The with a special thanks to Kelmend Zajazi and Natasha study and associated capacity development activities Ilieva-Acevska. benefited from the guidance and directions provided by Christine Kessides (Urban Practice Manager). An early version of the report was presented and discussed with representatives from local The team was led by Björn Philipp and comprised governments, local government associations, and Leif Jensen, Paula Restrepo Cadavid, and Kerem NALAS at a City-to-City Dialogue on Municipal Donmez. The regional research was led by Gabor Finance organized under the UPP by the World Peteri and comprised of local experts Anila Gjika, Bank Institute in Dubrovnik (Croatia) on June 20, Anto Bajo, Goran Rakic, Ljiljana Brdarevic, Marjan 2013. Preliminary findings and report results were Nikolov, Naida Carsimamovic, Natasa Obradivuc, discussed with the presidents and executive direc- and Osman Sadijak. tors of SEE local government associations at the NALAS General Assembly Meeting in Skopje (FYR The team has benefited tremendously from dis- Macedonia) on April 20, 2013. cussions with and comments from colleagues at vii Acronyms and Abbreviations BiH Bosnia and Herzegovina ECA Europe and Central Asia EU European Union FBiH Federation of Bosnia and Herzegovina LGA Local Government Association LGU Local Government Unit MFR Municipal Finance Review MoF Ministry of Finance MoLG Ministry of Local Government NALAS Network of Associations of Local Authorities of South East Europe PIT Personal Income Tax RS Republika Srpska SEE South East Europe UoM Union of Municipalities UPP Urban Partnership Program VAT Value Added Tax Vice President : Laura Tuck Country Director : Ellen Goldstein Sector Director : Laszlo Lovei Sector Manager : Sumila Gulyani Task Leader : Björn Philipp ix Executive Summary 1. BACKGROUND AND OBJECTIVE empirical information, since municipal finance time- series are not regularly available, let alone reliable The importance of local government finance has data for more sophisticated analytical approaches. increased over the last two decades. Decentralization Although Local Government Associations (LGAs) in the South East Europe (SEE) Region has given acknowledge the need for better municipal finance municipalities increased autonomy and shifted more data to strengthen their position in inter-govern- responsibilities for service delivery and resource mental negotiations, most lack comprehensive data management to local governments. An ever growing collection mechanisms or have limited access to share of public investment decisions is made by local already available information. authorities. Their relative weight in public spending and revenue raising has been increasing constantly The Western Balkans Municipal Finance Review over the last decade. On average, sub-national gov- (MFR) aims to address this shortcoming. Against ernment expenditure in the SEE region accounts this background, the objective of the analytical work for around 5 percent or more of GDP and are in under the SEE Municipal Finance Review aims to charge of around 15% of overall public expenditures. (i) contribute to improved understanding of local Municipalities, and the public utilities reporting to government management and finance in the SEE them, manage large and important infrastructure Region; and (ii) contribute to improving the qual- assets and need to secure financing for operation, ity and consistency of key municipal finance data maintenance, and capital renewal. for improved evidence based policy making. The analysis presented in this report comprises the first Despite local government’s important functions, attempt to review and analyze a regional set of the knowledge of municipal finance issues in SEE disaggregated sub-national finance data in the SEE is limited. Although reliable municipal finance Region. This analysis does not aim to be complete data is critical for fiscal forecasting, good financial and fully comprehensive, but to present options management, and to assist evidence-based planning for using disaggregated datasets for in-country and policy making, data is usually not accessible, and cross-country comparison. In particular, data or only to a limited degree. Overall data coverage collection designed and implemented under the and quality varies greatly among, and within, coun- MFR has the goal of leveraging the added value of tries. While information is generally available in disaggregated municipal finance data in a sub-region individual municipalities and at the national level, where generally, to date, municipal finance analysis there is a general lack of fiscal data across municipali- has been limited to aggregated, national datasets. ties. Regional comparisons and benchmarking are mostly absent, both within and across countries in This analysis is embedded in a regional capacity the region. Although the majority of municipalities building program to support local governments usually prepare some type of revenue and expendi- and their associations. Analytical activities and ture forecasting, not all attempt to make systematic technical assistance provided under the MFR were longer-term projections, partly due to the lack of co-financed and implemented jointly under the xi South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans World Bank-Austria Urban Partnership Program utility services such as water supply, sanitation, solid (UPP) in close collaboration between the ECA waste management, local roads, and basic social Sustainable Cities Initiative and the World Bank services. Health and education services are usually Institute’s Urban Practice. One of the key issues provided at a higher government level. However, the identified by national and municipal stakeholders level of decentralization and functional assignments consulted during preparation of UPP was the weak differ between and within the two groups of one financial management performance and low fiscal and multi-tier sub-national government systems. capacity of municipalities. MFR aims at increasing Serbia, Montenegro, and the Republika Srpska (RS) institutional capacity of local government for better in Bosnia and Herzegovina (BiH) have a more cen- municipal management in addition of demonstrat- tralized system in which the central government is ing the benefit and value added of disaggregated responsible for provision of health, education, and municipal finance data collection and analysis. social welfare. In Kosovo and Macedonia, munici- palities are allocated broader roles in primary and MFR has applied a ‘bottom-up’ approach for data secondary education, and primary health care. In collection and analysis. First, municipal finance addition, the role and weight of central and regional data sets were generated by involving LGAs and governments are very distinct in the different types local MFR experts in each country. A standardized of two and three tiered sub-national government framework on quantitative and qualitative dimen- systems in the SEE region. sions of municipal finance was developed, to ensure quality and comparisons of data collected in each But also the size and structure of municipalities country. Second, these disaggregated expenditure varies widely across the SEE region. There are con- and revenue data were linked to socio-economic siderable variations in the distribution of municipali- data sources, i.e, population size, density, urban/ ties by population size. The average population size rural classifications, location, national GDP and of local government units in one tier sub-national public expenditure, among others. The compre- government systems ranges between around 23,000 hensive combined datasets provide the platform inhabitants in the RS in BiH to close to 59,000 in for detailed analytical work, as well as support to Kosovo. On the other hand, countries with two-tier local governments and LGAs in achieving an in- sub-national governments like Albania and Croatia depth understanding of drivers and composition have a considerable amount of municipalities with of municipal finance. a population below 5,000 inhabitants. However, despite an overall large number of small munici- 2. DIFFERENT SHAPES OF palities in SEE, the highest population share con- DECENTRALIZATION IN THE centrates in medium and large cities (see Figure 1). WESTERN BALKANS Municipalities in SEE are also very diverse in terms of population density. Variations in the group of larg- The specific institutional framework and admin- est municipalities with more than 200,000 inhabit- istrative structure differs—despite the common ants range from more than 1,600 inhabitants per km2 history and overall perspective of European Union (Serbia) to less than 200 per km2 in the RS. Within (EU) accession shared by many countries in the countries, divergences in densities are particularly SEE region. However, after more than a decade-long striking in Croatia, Kosovo and Montenegro. decentralization reforms in the region, the overall institutional framework has been put in place and Overall, capital cities dominate the local govern- stabilized in the seven Western Balkan countries. In ment system in SEE. Despite the seemingly balanced general, SEE is characterized by a single tier local population distribution, there is a spatial concen- government system with relatively large size munici- tration in metropolitan areas. From the 7 countries palities; only Albania, Bosnia and Herzegovina, and in SEE included in the analysis, 4 have more than Croatia have two and three tier sub-national govern- 15% of their population living in the capital city. ment systems. Typically, municipalities provide core Most accentuated in Montenegro, 30% of the total xii Executive Summary Figure 1: Distribution of municipalities across different sizes (left) and population distribution across different size municipalities (right) 100% 100% % of citiesin each population range 90% 90% % population per size range 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Albania BiH Croatia Kosovo Macedonia Montenegro Serbia Albania BiH Croatia Kosovo Macedonia Montenegro Serbia Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 population resides in the capital city. Capital cities Overall, grants, shared revenues and transfers domi- generally also generate the largest share of GDP. nate local revenue composition. Grants and transfers For example, in Macedonia, Skopje accounts for account for a particularly large share of sub-national about 25% of the total population and contributes budgets in Albania and Kosovo. Only in the RS and to around 45.6% of the country’s GDP. Montenegro, own-source revenues account for more than half of sub-national budgets (see Figure 2; left 3. COMPOSITION AND side). However, weight and composition of own- STRUCTURE OF LOCAL source revenues varies across and within countries. GOVERNMENT FINANCE IN THE In Albania and the RS, the largest share of own source revenue originates from taxes, while Croatia, Kosovo WESTERN BALKANS and Macedonia have a more diversified revenue Local government funding and expenditure struc- composition which includes revenue from rents and tures are very diverse. Variations in the revenue capital. During the period of analysis, total local composition across the SEE region are large, with revenues have increased in Albania, Kosovo and the share of own source revenues ranging between Macedonia, but took a hit in the other countries (see 15% in Kosovo to 82% in the RS. Variations in the Figure 2; right side). The decline of local revenues was sub-national expenditure structure are caused by only partially compensated by national transfers, since the different functional assignments—regardless central government budgets also contracted during of general government spending—but are also a the same period, and was particularly contrasted in reflection of the diversity of intergovernmental Croatia, Montenegro and Serbia. In Macedonia, vari- arrangements. For example, while the overall public ous earmarked transfers from line ministries dominate sector in Albania and Kosovo is markedly smaller local budgets. Kosovo municipalities benefitted from than in the other countries, the importance of sub- an overall increase of close to 92% during 2008–2011, national government varies almost by a factor four: driven largely by growing grants and transfers such sub-national governments in Albania account for as in Macedonia (40%). 8.8% of total public expenditures, while in Kosovo, almost 49% of all public expenditures are handled Overall, local government expenditures in SEE by sub-national governments. range between of 2–8% of GDP. At that level, xiii South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 2: Composition and evolution of local government revenues Percentage General Revenue by source – 2011 Regional Comparison - Evolution of Total Revenue (2010 Federation of BiH) 2008 – 2011 (2010 for BiH RS and Serbia) 100% 100% 90% 75% 80% 70% 50% 60% 25% 50% 40% 0% 30% –25% 20% 10% –50% 0% –75% Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro –100% 2008 2009 2010 2011 Own source Shared taxes Grants and Transfers Albania Federation of BiH BiH RS Croatia Borrowing Other Kosovo Macedonia Montenegro Serbia countries in the SEE region are generally in line with Kosovo: here, local expenditures are less than half of comparable countries in the EU, but below the EU27 the average in Montenegro. However, by far the low- average of 14%. The shape of decentralization influ- est level of per capita spending is observed in Albania. ences general government expenditure. For example, On average, Albanian municipalities only spend 10 Bosnia and Herzegovina pays a premium for the more percent of the per capita average in Montenegrin complex institutional structure inherited as a result of municipalities. While this may be a reflection of the the Dayton Peace Agreement in 1995: although BiH’s multi-tier sub-national government system, equally GDP is below the regional average, public sector to the FBiH, it also reflects the generally low level of spending reaches 49.2 percent. However, on average, local expenditure as share of GDP in Albania. the overall weight of local government spending has decreased since 2008, less two exceptions: local General public services typically dominate local expenditures increased in Macedonia from 4.3% of expenditure. This category, which generally covers GDP in 2008 to 5.6% in 2011; and in Kosovo from services of elected organs and the local administra- 5.2% to 8.4% in the same period. This reverse trend, tion, accounts for more than half of sub-national however, is largely the result of education spending expenditure in Albania and Montenegro, and close which was assigned to municipalities in both coun- to 50% in the FBiH. Only in Kosovo and Macedonia, tries. The weight of local government spending has education accounts for the largest share of local decreased the sharpest in Montenegro, from 5.5% of expenditure. In both countries, municipalities have GDP in 2008 to 3.9% in 2011. full responsibility for primary and secondary educa- tion, including teachers’ salaries and wages, which However, per capita local government expenditure make up the lion share of local spending with 48.7% in the Western Balkans is low. On average, local in Macedonia and 42.2% in Kosovo. governments in the SEE region spend EUR288 per capita, compared to EUR 2986.3 in the EU27. Croatia Housing, communal services and economic affairs and Montenegro are an exception with an average account for the largest share of urban services. of around EUR500 per capita. Local government Typically, local governments are responsible for hous- spending is much lower in the Federation of BiH and ing development and maintenance, water supply and xiv Executive Summary public street lighting, accounted under ‘housing and system, it is important to understand the nature of communal services’. This category constitutes a sig- grants and transfers, in addition to their size. For nificant part of local budgets in Macedonia (19.6%), example, while sub-national governments’ expen- Croatia (17.2%), the Federation of BiH (17.7%), diture in Albania are funded to a large extent by and Montenegro (14.4). Public transportation is central transfers (69% in 2011), the funds are almost the largest single item accounted for in ‘economic fully provided as unconditional grants which leave affairs’. This sub-sector covers road building and room for spending decisions at the sub-national maintenance, grants provided to operation of roads, governments’ discretion. On the other hand, while railway and other public transportation systems. If grant dependency per the share of local govern- local governments are directly involved in tourism, ment financing seems much lower in Croatia and restaurant and hotel businesses, then they are also Montenegro, the grant ‘picture’ is also much more accounted here. Economic affairs are present in all complex with a larger share of ear-marked grants. local budgets, but they have the highest shares in Although ear-marked grants are used to ensure Albania (26.8%) and Croatia (15.2%). Environmental standards and quality of core public services like protection services, such as municipal solid waste education and health, they leave no or little financial management and sewage treatment are less visible discretion to sub-national governments. among the local government expenditures, because they are either underdeveloped or reported by public Overall, sub-national borrowing remains a minor utilities responsible for communal service provision. source of revenue in SEE. Total accumulated local government debt is not significant, at less than 1% of Grants and transfers form an important part of GDP, even in countries with more actively borrow- sub-national budgets. However, the size and nature ing sub-national governments, such as Montenegro of ‘grant dependency’ varies significantly. Grants and Serbia. Among others, this reflects the strict and transfers account for more than 50% in Albania, sub-national borrowing regulation across the SEE Kosovo and Macedonia. Apart from Montenegro, the region. All countries have borrowing laws which share of grants in total has increased in all countries. control local government debt and limit sub-national In Albania and Kosovo, the grant share was increased borrowing by revenue and outstanding debt. In all significantly over the considered period, reaching countries, sub national governments need consent 69% and 85% respectively in 2011, while Macedonia or approval from the central government in order saw a much more moderate development. In to take a loan, guarantee or issue a bond. However, Montenegro, on the other hand, grants and transfers the role of municipal credit would be expected to from the central government decreased by more than increase with greater financing needs for local infra- 10 percent from 26.4 in 2008 to 15.4 in 2010; and only structure investments and more restrictive central recovered somewhat again in 2011, stabilizing at 20.8 government budgets. percent. The relative little importance of grants and transfers in Croatia and Montenegro should be seen Imbalances in sub-national debt may become a in the context of the overall funding profile: as the cause of future fiscal stress. The distribution of debt only two countries in the SEE region, Croatia and across sub-national borrowers is very uneven across Montenegro benefit significantly from shared taxes municipalities and may pose a serious fiscal risk for as funding source. In Croatia, shared taxes are the some highly indebted local governments. In Albania, most important funding source, covering about 50% mostly larger municipalities and local governments of expenditures. Although shared taxes play a much in less developed regions borrow. The concentration less important role in Montenegro, they still cover of debt in weaker municipalities is of particular con- 12–13% of the sub-national government funding. cern. In Montenegro, some of the poorer municipali- ties are more actively acquiring loan funds, despite However, the type of grants defines dependency the heavily regulated local government borrowing beyond their share in local budgets. To fully procedures. Annual debt repayment in percentage understand and assess the inter-governmental fiscal of the total current expenditures has reached 53.6% xv South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans in Montenegro where local debt is concentrated in more than 200,000 inhabitants represent around the coastal cities and the local governments in the 45% of the total own source revenues of local Northern region. These municipalities cannot cope governments, while they are home to only around with the high debt repayment burden, so a special 18% of the population. Similar results are found in financial recovery plan had to be signed with the Kosovo. In Albania, Macedonia and Montenegro, Ministry of Finance. on the other hand, medium size municipalities concentrate most of the countries’ own-source revenues—despite the lower per capita own-source 4. VARIATION AND DISPARITIES revenues in those smaller municipalities. This is ACROSS MUNICIPALITIES WITHIN mainly the result of the large number of medium COUNTRIES size municipalities in this latter group of countries. At the same time, smaller municipalities seem to The composition of local revenues differs across underutilize fees and charges in Macedonia and but also within the SEE countries. A closer review of Montenegro. Overall, findings suggest that par- the disaggregated data reveals significant variations ticularly municipalities with a population smaller across local governments within countries. Factors than 10,000 have significant revenue enhancement such as the local economic base and property values, potential. among others, determine the room for collecting property related taxes or business fees. For instance, Expenditure variations are largely caused by the property values tend to be higher in larger munici- service function concentration in geographical palities when compared to small municipalities. centers. For example, in Albania, half of the total Overall, bigger municipalities tend to have higher local public education spending is made in Tirana. per capita tax revenues compared to municipalities In the Federation of BiH, water services (housing with a smaller population. and community amenities), health care, recreational, sports and culture, and social protection services Total revenue trends suggest large variations across are concentrated in Sarajevo and the cantonal main municipalities of different size in Albania and cities. However, asymmetric functional assignments Montenegro. While, overall, down- and upward amplify variations. For example, during the decen- trends in total local revenues show comparably tralization process in Croatia, local government func- similar patterns in the municipalities of BiH, Croatia, tions were devolved to the city level very selectively. Kosovo, and Macedonia; trends vary significantly Only the 32 financially stronger municipalities were for municipalities of different size in Albania and authorized to take over additional responsibilities, Montenegro. In Albania, we find overall stagnation of such as education, housing and community ameni- total revenues across municipalities after an increase ties, reflected in a higher expenditure share of those from 2009–2010. However, municipalities with a categories in these selected municipalities. population between 100,000–200,000, MFR data indicate a strong decline in total local revenues. In General public service expenditure per capita Montenegro, on the other hand, smaller municipali- tends to decrease with larger population size. By ties have increased their total revenues; in particular and large, findings suggest that the share of general municipalities with less than 5,000 inhabitants reg- public service expenditure of total local expenditure istered a steep increase in total revenues after 2009. increases in smaller municipalities. In Kosovo, for example, municipalities with less than 5,000 inhabit- However, own-source revenues are not necessarily ants have much higher spending than medium and concentrated in the biggest municipalities. While large municipalities. A similar patter can be found in some countries, e.g., Croatia and Kosovo, we find in the Federation of BiH. This might be a reflection a high concentration of own-source revenues in the of economies of scale, since smaller municipalities larger centers, this is not a consistent pattern across generally face higher per capita fixed costs for provi- the SEE Region. In Croatia, municipalities with sion of a minimum service level. xvi Executive Summary Figure 3: Geographic concentration of total revenues by city size Croatia Macedonia 7.00E+09 1.20E+10 6.00E+09 1.00E+10 5.00E+09 8.00E+09 4.00E+09 MKD HRK 6.00E+09 3.00E+09 4.00E+09 2.00E+09 1.00E+09 2.00E+09 0.00E+00 0.00E+00 less than 5000 5000 to –10000 10000–40000 40000–100000 100000–200000 more than less than 5000 5000 to –10000 10000–40000 40000–100000 100000–200000 more than 200000 200000 Own source Shared taxes Grants and Transfers Own source Shared taxes Grants and Transfers Borrowing Other Borrowing Other Overall, urban areas generate a higher share increase in own-source revenues. It is possible that of own-source revenues. Among the countries capital sales in rural areas have a higher effect on included in the MFR, Macedonian municipalities their total revenues when compared to urban areas. are classified as urban or rural which allows for additional analysis along this categorization. There Rural municipalities tend to have higher per capita are considerable variations between urban and rural expenditure than urban municipalities—particu- municipalities in the composition of total revenues, larly in the functional categories of environmental but also the share of own source revenues. Revenue trends in Macedonia suggest rural municipalities outperform their urban peers in mobilizing rev- enues. During 2008–2011, rural municipalities’ total Figure 4: Variations on per capita expenditures revenues grew faster than urban municipalities in General Public Services within municipality total revenues. However, a closer review of the data ranges – Kosovo reveals that rural municipalities’ better performance 400 in total revenue mobilization was largely the result of increasing grants and transfers from the central 300 government. In addition to a higher share of own- source revenues, urban municipalities also generate 200 a higher percentage of own-source revenues from taxes, fees and charges. More surprisingly, capital 100 revenues are significantly higher in rural areas. However, this may suggest underused revenue 0 enhancement potential from taxes, fees and charges less than 5000 5000 to –10000 10000–40000 40000–100000 100000–200000 more than 200000 in rural municipalities and might also be a reflection of recent asset sales: starting in 2009, Macedonian municipalities gained control over land transac- tions which resulted in numerous asset sales and an xvii South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans protection, education, health and general public Capital expenditures are concentrated in cities and services. Housing is the only expenditure item in larger municipalities. Although no major conclusions which urban municipalities spend more per capita can be made from capital expenditure data limited to compared to rural municipalities. This might be a one year period and should be interpreted carefully, explained by larger economies of scale in urban some observations can be noted. Overall, capital cities municipalities. However, although urban areas con- and other urban centers dominate capital spending centrate a larger share of total capital expenditure, across the SEE region, particularly in Serbia (Belgrade: per capita capital expenditures are slightly larger in 68%), Macedonia (Skopje: 59%) and Montenegro Macedonia’s rural municipalities. (Podgorica: 45%). In Albania, BiH and Croatia, mid- size municipalities account for the highest share of Urban density partly explains disparities among total capital expenditure. This concentration is the local governments’ expenditures in the SEE Region. result of the large number of smaller and medium size Density is relevant for the unitary cost of provision municipalities in these countries. However, despite of selected municipal services, such as water supply the similar share of total capital expenditure in these and sanitation, solid waste management, and public countries, the share of capital expenditure to total transport, among others. Based on the MFR analysis, local expenditure varies significantly: from only 17% density effects can be observed predominantly on in Croatia to over 33% in Albania. environmental protection expenditure. In the case of Croatia and Macedonia, density correlates with However, there are large variations between total a significant reduction in per capita expenditures and per capita capital expenditures. In Albania, for for environment protection. These results provide example, the biggest share of total capital expenditures evidence to support the assumption that denser is concentrated in municipalities with 10,000–40,000 municipalities tend to have lower per capita cost for inhabitants. However, the largest mean capital expen- environmental protection expenditures. diture can be found in municipalities ranging with more than 200,000 inhabitants (Durres and Tirana). 5. CONCENTRATION IN CAPITAL The per capita capital expenditures analysis reveals a more or less homogenous national scenario, although INVESTMENT EXPENDITURE AND Durres and Tirana—the two biggest cities in Albania— THE ROLE OF CAPITAL CITIES have the highest per capita capital expenditure. Generally, the scope of local capital expenditures Overall, capital cities dominate total sub-national varies with the level of economic development. expenditures. Spatial concentration of local expen- Capital expenditures constitute a major part of ditures in the capital cities is especially prevalent local expenditure in the SEE region. Municipalities Albania (Tirana), Bosnia and Herzegovina (Sarajevo) spend more than 25% of their available funds on and Serbia (Belgrade), where the share of local capital investments, with the exception of Croatia expenditures significantly exceeds the population (17%). Overall, local capital expenditures per capita share living in the capital. In addition, in all of the vary between EUR48 in Albania and EUR459 in SEE countries, except Albania and Croatia, there is Croatia, or 2.2% and 6.8% percent of GDP. Overall, a disproportionate concentration of capital expendi- variations in local capital expenditure levels are in tures in capital cities (see Figure 5). Macedonia and line with the differences in economic development. Serbia are exceptional cases even among countries However, Kosovo is a noteworthy exception: per with the highest concentration. In Belgrade, local capita local capital expenditure is above the aver- spending is more than two times higher than the age at 8.4% percent. Functional assignments and share of local residents: 44.6% of overall local govern- funding mechanisms have a large influence. Local ment budget is spent in the capital city, while only governments in more decentralized countries spend 20% of the population lives in Belgrade. Skopje, in more on capital expenditures than their peers in less Macedonia, concentrates more than 70% of total decentralized countries. capital expenditure in the country. The concentration xviii Executive Summary Figure 5: Concentration of expenditures and population in the capital city 80% 70% 60% 50% 40% 30% 20% 10% 0% Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro Serbia Share of total expenditure Share of population Share of capital expenditure of expenditures in capital cities might be partly have been hit earlier by the fiscal restrictions and explained by the fact that most of these cities are contracting economic outputs. Already in 2009, local assigned additional responsibilities’ for public ser- expenditures dropped significantly in Montenegro vice delivery. (–18%) and in the Republika Srpska (–8.6%); and decreased slightly in Serbia (–3.7%) compared to the At the same time, capitals benefit from larger prop- previous year. In comparison, the crisis was delayed erty tax revenues. Across the SEE region, capital cit- and less drastic in, e.g., Kosovo and Macedonia, but ies collect the lion share of property related revenues, also Albania and Croatia. Particularly in Kosovo including property taxes and fees, and communal and Macedonia, local government expenditures charges levied on the property of businesses or resi- continued to increase in real terms throughout this dents. Overall, the concentration of property related period, albeit with a slower pace. In Albania, local revenues in the capital city exceeds both the share expenditures were still growing in 2009; and in of population and the budget. They benefit from the Croatia, local expenditures did not decline—possibly disproportionate concentration of the tax base, but due to the national elections in 2010 and increased also typically higher property values. EU funding that was made available to local authori- ties. However, cutbacks were more serious later in 6. KEY TRENDS, DISPARITIES AND 2010 (–9.8%). These trends are also reflected by the changes in local expenditures as share of in GDP. INEQUALITY ACROSS WESTERN BALKANS’ MUNICIPALITIES Large cities didn’t necessarily do better than their smaller peers during 2008–2011. For example, Local governments in SEE felt the downturn at dif- Prishtina (Kosovo) and Tirana (Albania) seem to ferent points in time. The 2008 economic and finan- have been hit harder by the crisis than other local cial crisis had a strong influence on public finances governments in their countries, while Zagreb since 2008 and continues affecting fiscal policies in (Croatia), Podgorica (Montenegro) and Belgrade the SEE region and beyond. From the MFR data, (Serbia) appear to have done better during the first covering 2008–2011, it is not possible to differenti- year of the crisis. Partly, this can be explained by ate the effects of the crisis from general structural the local revenue structure: Croatia and Serbia rely trends. However, an analysis of the trends over this on shared revenues which automatically delayed period suggest that some local governments might the decline in resources. In Montenegro, however, xix South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans coastal municipalities, which dominate total local Poorer municipalities could not consistently catch- expenditures, had a more sudden fall in property up with their richer peers. Per capita expenditure related tax revenues and tried to compensate the trends reveal increased disparities between rich and revenue decrease with asset sales. Poorer munici- poor municipalities in some countries, and conver- palities, on the other hand, were forced to increase gence in others. In Kosovo, poorest municipalities borrowing. Macedonian municipalities were able have achieved to close the gap to richer municipali- to maintain their own sources revenues by follow- ties in their per capita expenditures. In Macedonia, ing different policies in the urban and the rural on the other hand, poorer municipalities have not municipalities. The immediate reaction of the rural achieved to close the gap. Here, per capita expendi- municipalities was to increase local fees. They ture increased overall in parallel in the richest and also started to borrow from a very low base. The poorest municipalities, although with a slightly urban municipalities and Skopje had the option to diverging trend. Worryingly, the spending difference sell their assets. They also increased the revenues between poor and rich municipalities grew con- from loans. siderably in Montenegro (see Figure 6). In Croatia, disparities were maintained during 2008 and 2011, Local revenues in Albania and Kosovo were and Albania experienced a slight increase in the gap stabilized by transfers and grants. In Kosovo, between the poorest and the richest after 2010. local government revenues recovered relatively soon after their fall in 2010. Prishtina seem to have Spending patterns differ remarkably in munici- been hit most by the economic downturn, causing palities with high and low per capita expenditure. a sharp fall in own source revenues, while local In Kosovo, while there is convergence in per capita governments in the rest of the country were able to expenditures, poor and rich municipalities have very keep previous levels of own source revenues. But different spending priorities. Poor municipalities local government budgets in Kosovo are mostly increased considerably their expenditures on health financed by grants, which were stabilized at a lower and general public services while, during the same level of growth, but still continued to increase in period, rich municipalities increased their expen- 2010–2011. Albanian sub-national governments ditures on environmental protection and housing. faced restrictions with some delay since uncondi- tional transfers remained stable in 2009 and later, Per capita own-source revenue inequality increases the cuts were spread equally among the different in Albania, Kosovo, Montenegro, and the RS. types of local governments. Findings suggest a visible growth of per capita Figure 6: Per-capita expenditure trends in poorest (1st Quintile) and richest (5th Quintile) municipalities for Kosovo (left) and Montenegro (right) (reference year: 2008) 140 1,600 Mean per capita expenditure per Mean per capita expenditure per 120 1,400 100 1,200 revenue quintile revenue quintile 1,000 80 800 60 600 40 400 20 200 0 0 2008 2009 2010 2011 2007 2008 2009 2010 2011 1 5 xx Executive Summary own-source revenue inequality during the period 7. MUNICIPAL FINANCE DATA of analysis, measured by the Gini coefficient meth- COLLECTION AND ANALYSIS IN odology. In Croatia and Macedonia, results suggest a slight reduction in per capita own-source rev- SEE: DIRECTIONS FOR A POSSIBLE enue inequalities during 2008 and 2011. Total per WAY FORWARD capita revenue inequalities grew in RS, Kosovo and Montenegro, but were reduced in Albania, Croatia This report has demonstrated the increasing rela- and Macedonia. However, for all years and coun- tive weight and role of sub-national governments tries, except for the RS in 2009, the level of per capita in the SEE region. Greater fiscal authority and revenue inequality is greater when computing own responsibilities drive a need for improved planning, source revenues as opposed to general revenues, budgeting and financial management—both at the suggesting that equalization grants are working. local and central level. Reliable and high quality municipal finance data is critical to better under- Overall, the general grant transfers seem to meet stand and improve local government management equalization objectives across the SEE region. and finance in the SEE Region. In this context, this Generally, local governments with the highest report has aimed to address identified capacity amount of per capita transfer received 1.5–3.8 more issues by the creation of a regional dataset for disag- than the average in the respective group of cities or gregated municipal finance analysis with the goal to municipalities in the majority of SEE countries. In contribute to improving the quality and consistency addition, the level of grant per capita is decreasing of key municipal finance data for improved evidence by size of municipalities, reflecting the efforts to based policy making. Although limited in scope and compensate for low fiscal capacity/higher expendi- expandable in depth, MFR has piloted disaggregated ture needs in the smaller, often rural municipalities, data collection and analysis for in-country analyses which may not be in a position to afford service and regional comparison in SEE. delivery without raising tax efforts excessively. MFR has established a regional municipal finance Per Capita Expenditure, own sources, and grant database available for further analysis—with the transfers by size of municipalities vary signifi- support from regional and local experts and the cantly across countries. In Albania and Kosovo, Local Government Associations in the participating the per capita grant is by all measures the dominant entities. However, the potential use and application fund source in smaller municipalities, while in of disaggregated data go beyond the initial analysis Montenegro and Macedonia own fiscal efforts per possible within the scope and limitations of the MFR. capita are more important, although not equal to For example, compared to aggregated national data- grants. Expenditure variations and levels, i.e., expen- sets, disaggregated sub-national finance data allows diture needs, by size of municipalities vary across for benchmarking across municipalities. Municipal the countries. In Kosovo, the per capita expenditure current and capital expenditures analyzed together is decreasing by size of municipalities, implying that with service performance indicators and statistics per capita spending in the largest municipalities may provide opportunities for comparing facility amount to less than 50% of per capita spending in the unit costs, but also to reviewing variations in service smallest municipalities. In Albania and Macedonia, delivery outcomes against different expenditure on the other hand, while per capita expenditure needs; spending- and revenue trends. do not differ significantly across the various sizes of municipalities in these two countries, a minor Going forward, a number of questions for follow- increase by size is observed in Albania, with expen- up actions may be considered. These include, diture of +20% in the largest municipalities. Finally, among others: How to sustain regional and country- in Croatia, a somewhat similar pattern to Albania is specific datasets and analytical capacity within SEE observed, although the expenditure is much higher countries for continued analysis, including outreach in the larger municipalities than in smaller. to national authorities on data and analysis? How xxi South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans to establish further capacity of local associations overall performance of a municipality. This requires in each country to maintain, expand and update public access to selected information, or to indica- the database? How would a regional database be tors, depending on the level of available municipal maintained, and by whom? How to broaden the finance data. Publicly accessible databases can also concept of a database and analytical review and offer search options, such as aggregation by regions, include other countries in the ECA region? A number comparison of local governments, or the averages of issues would need to be addressed to respond to for the a subset of municipalities (e.g. similar size, these questions with the goal to (i) make municipal administrative status). finance data collection regular; and (ii) develop a user friendly system of information management. However, data management capacity needs to be strengthened. Generally, local government associa- The model templates and analysis piloted under tions are not prepared for the complex task of man- MFR can be used for further scaling-up. In all aging municipal finance data and present the gener- countries in the SEE region, disaggregated financial ated information in an accessible and user-friendly reports on local government expenditures and rev- format. In the past, LGAs have not had a business enues are available at the Ministry of Finance or at model for developing and operating municipal the Treasury. The only exception is the Federation of finance data and typically relied on donor support. BiH, where the local government association with For example, the LGAs in Albania, Montenegro, and the support from the MFR has launched a survey to the Federation of BiH have expressed their interest collect municipal fiscal data. The survey format was in leveraging donor financial assistance for regular similar to the financial reports used by local govern- municipal finance data management. At the regional ments. The approach for data collection and clean-up level, if demonstrated demand exists, NALAS would applied under MFR can be replicated in the future. be a natural host for a regional municipal finance database. However, related costs of data collection Public access to data may contribute to enhanced and maintenance need to be planned and budgeted transparency. Disaggregated municipal finance upfront to ensure sustainability. Since local gov- data can be used for various purposes. Among ernment associations in the target countries are in others, the goal of municipal finance data and different positions with regard to their mandates, analysis is to inform local government officials, organizational, and staff capacities, business models investors, media, civil society organizations and for long term sustainability necessarily will have to the general public about the financial position and be different. xxii 1 Introduction 1.1 BACKGROUND Local governments in South-East Europe (SEE) have a history of assigned responsibilities for delivering municipal infrastructure and services, while national governments provided financing and approved local investment projects. More recently, decentralization has given municipalities increased autonomy and shifted more responsibilities for service delivery and resource management to local governments. Fiscal transfers continued to constitute the larger portion of local gov- ernment revenues. However, with the fiscal and financial crisis, these resources never commensurate with the rising needs. At the same time, an ever growing share of public investment decisions is made by local authorities. Their relative weight in public spending has been increas- ing constantly over the last decade. Municipalities, and the public utili- ties reporting to them, manage large and important infrastructure assets and need to secure financing for operation, maintenance, and capital renewal. In parallel, sub-sovereign borrowing restrictions have been eased widely throughout the region to provide additional financing means to local governments. A growing number of municipalities aim at improving creditworthiness and municipal financial management performance to gain access to commercial funding sources. Against this background, the way municipalities raise, allocate, and manage their financial resources becomes increasingly important. Greater fiscal authority and responsibilities fuel the need for improved budgeting and planning—both at the local and central level. Reliable municipal finance data is critical for fiscal forecasting, good financial management, and to assist evidence-based planning and policy mak- ing. Furthermore, commercial lenders will assess municipal creditwor- thiness based on the available financial data. Typically, in SEE countries, key municipal finance functions include revenue and expenditure management, cash and debt management, local budgeting, and accounting, but increasingly, municipal asset management and capital investment planning. However, despite those 1 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans important functions, the knowledge of municipal Urban Partnership Program (UPP) in December finance issues in SEE is limited. Data is usually not 2010 was the weak financial management perfor- publicly accessible, or only to a limited degree, and mance and low fiscal capacity of municipalities. data coverage and quality varies greatly among, Local authorities perceive municipal financial and within, countries. Recent surveys conducted management predominantly as centrally imposed on behalf of the Network of Associations of Local planning, reporting and control systems, rather than Authorities in South-East Europe (NALAS) sug- as a management and policy making tool. Mayors gested that expenditure data is available only at very and municipal councilors have little incentives to different levels of specification, even at the national, demand better municipal financial performance, aggregate level. On the other hand, revenue data, in since comparative data is rarely available, or made particular own source revenues, are usually reported accessible, and the link between improved financial in great detail, although classification differs signifi- performance, higher service level outcomes, and cantly and the information on shared revenues and better access to market-based capital investment intergovernmental transfers is limited.1 Given the funding remains blurred. Although LGAs acknowl- tight sub-sovereign borrowing controls, information edge the need for better municipal finance data to on municipal debt is generally good. Usually, local strengthen their position in inter-governmental government fiscal data are collected by the Ministry negotiations, most lack comprehensive data collec- of Finance (MoF) and/ or the Treasury, who control tion mechanisms or have limited access to already access to the disaggregated data. In some countries, available information. alternative sources of municipal data exist, often generated with support from donors and/or col- 1.2  OBJECTIVE AND lected through the Local Government Association (LGA). In general, however, while information is METHODOLOGY available in individual municipalities and at the national level, there is a lack of fiscal data across Against this background, the objective of the analyti- municipalities. cal work under the SEE Municipal Finance Review aims to (i) contribute to improved understanding of As a result, municipal budgeting and fiscal manage- local government management and finance in the ment practices often remain based on judgment. SEE Region; and (ii) contribute to improving the Central government capital and current grant allo- quality and consistency of key municipal finance cations often lack the datasets required to establish data for improved evidence based policy making. In consistent and formula-based models. Regional com- the short-term, the report would inform government parisons and benchmarking are mostly absent, both authorities in the decision making process through within and across countries in the region. Although the provision of empirical data for evidence-based the majority of municipalities usually prepare some policy making. In the medium-to-long term, access type of revenue and expenditure forecasting, not all to municipal finance information would contribute attempt to make systematic longer-term projections, to increasing transparency and accountability of partly due to the lack of empirical information, since local governments, improving revenue collection municipal finance time-series are not regularly avail- and expenditure performance, optimizing budget able, let alone reliable data for more sophisticated allocation procedures, and strengthen local authori- analytical approaches. ties’ role and position in intergovernmental fiscal considerations and negotiations in the SEE countries. One of the key issues identified by national and municipal stakeholders consulted during scoping In particular, data collection designed and imple- missions in preparation of the World Bank-Austria mented under the MFR has the goal of leveraging the added value of disaggregated municipal finance data. Generally, to date, municipal finance analysis in the 1 NALAS (2012), Report of Fiscal Decentralization Indicators for South-East Europe 2006–2011. SEE region has been limited to aggregated, national 2 Introduction datasets.2 In this report, municipal data are estab- possible, this report explains the selected approach lished in two steps: First, municipal finance data sets and underlying rationale. Detailed methodological are being generated ‘bottom-up’ by involving Local guidance was provided to the LGAs and summa- Government Associations and local MFR experts in rized in the country-specific reports. A summary each country. A standardized framework on quantita- overview with further details and explanation on tive and qualitative dimensions of municipal finance the data collection approach is attached in Annex 3. was developed, to ensure quality and comparisons of data collected in each country. The framework National governments, during fiscal planning, usu- unbundles finance data to a larger extent than what ally assess local governments as one, compact unit of is normally available in national datasets, with expen- general government finances. In some cases, policy ditures and revenues distributed by type and size of makers assess impacts of regulatory changes based cities and other municipalities, and by establishing on simulations by different types of municipalities additional measures on categories of revenues and or geographic regions. However, usually the results borrowing. Second, these disaggregated expenditure are not shared with a broader group of stakeholders and revenue data are then linked to socio-economic who contribute to the policy dialogue on local gov- data sources, i.e, population, urban/rural classifica- ernment finance. MFR has developed an approach tions, area, GDP, and national expenditure, revenue, to establish disaggregated municipal finance data- debt, and capital investments. The comprehensive sets and equip stakeholders with reliable fiscal combined datasets provide the platform for detailed information to contribute to this policy dialogue. analytical work, as well as support to local govern- In particular, the tools and findings of MFR would ments and LGAs in achieving an in-depth under- be deemed to help strengthening LGAs planning standing of drivers and composition of municipal capacity and better position them in the emerging finance, which in turn may lead to improved fiscal policy dialogue on local government finance—both performance. The datasets also allow for regional at the country-specific and regional level. comparison among the SEE countries participating in the MFR. Where data in time series is available, this Main findings of the MFR are presented in this report will also identify the impact of the economic report. Following this introductory chapter, Chapter downturn in the period of 2008–2011 and assess the 2 will provide an overview of the decentralization fiscal responses developed in the region. framework in the SEE Region, including on the administrative and political structure of sub-national The analysis presented in this report comprises the governments, their population size and distribu- first attempt to review and analyze a regional set of tion, and the service functions assigned to local disaggregated sub-national finance data, including governments. Chapter 3 explains in more detail the from municipalities, cities, counties and other inter- local government finance framework. This includes mediary government-levels from various countries. an overview of the structure and composition of This analysis does not aim to be complete and fully sub-national finances, in particular (i) revenue and comprehensive, but to present options for using expenditure assignments; (ii) transfers and inter- disaggregated datasets for in-country and cross- governmental fiscal relations; and (iii) the evolving country comparison. Different reporting standards framework and realities of sub-national borrowing and practices across the region required reclassi- and debt. Chapter 4 provides a summary of the key fication and re-coding of data during the process trends and findings from the cross-country, regional of data cleaning. For the data analysis, all values analysis, complemented by detailed analysis of the were adjusted for inflation and correspond to 2008 disaggregated datasets, where available. Finally, prices. Wherever different interpretations would be Chapter 5 summarizes conclusions and provides some recommendations for a possible way forward. 2 NALAS 2012 3 2 Overview: Decentralization in the Western Balkans 2.1  ADMINISTRATIVE STRUCTURE AND FUNCTIONS OF SUB-NATIONAL GOVERNMENTS South Eastern European countries have reached the second stage of development in establishing local governments. After more than a decade-long decentralization reforms and based on the local self-gov- ernment traditions in the region, the overall institutional framework has been put in place and stabilized in all seven countries. However, despite the common history and overall perspective of European Union (EU) accession shared by many countries in the SEE region, the specific institutional framework and administrative structure differs significantly. The following summary provides a brief overview of the main features in each of the seven countries. SEE countries have different sub-national government structures and service responsibilities. However, in general, the region is character- ized by a single tier local government system with relatively large size municipalities. Most of the SEE countries, i.e., Serbia, Kosovo, Macedonia, and Montenegro, inherited large political/administrative entities which were kept, or partially restored, e.g. in Macedonia. In those countries, municipalities as the lowest level of sub-national gov- ernment represent the core functional units. Only Albania, Bosnia and Herzegovina (BiH), and Croatia have two and three tier sub-national governments: counties, communes and municipalities in Albania; entities, cantons and municipalities in Bosnia and Herzegovina; and counties, county centers, large cities and municipalities in Croatia. Generally, municipalities provide core utility services such as water supply, sanitation, solid waste management, local roads, and basic social services. Health and education services are usually provided at a higher government level. However, the level of decentralization and functional assignments differ between and within the two groups of one and multi-tier sub-national government systems. Serbia, and Montenegro, and Republika Srpska (RS) in BiH have a more centralized system in which the central government is responsible for provision 5 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans of health, education, and social welfare. In Kosovo been organized as coordinating bodies with very and Macedonia, municipalities are allocated broader little exclusive responsibilities for preparing regional roles in primary and secondary education, and pri- policies and implementing them in harmonization mary health care.3 with national policies such as regional and urban planning, regional transportation services and envi- The two and three tiered sub national government ronmental protection. Education, including curricula systems in the SEE region also feature different development and staffing, and health care services levels of decentralization. In Albania, the central are essentially provided by the central government. government plays an important role in the field of Although the organic law on local government has education and health, but quarks have intermediary established a sound foundation for the assignment of roles between central and local governments, such as expenditure responsibilities, the initiative for further social assistance and environmental protection. In the specifying responsibilities has essentially been left Federation of BiH, cantons, as regional governments, to the decisions of line ministries which, in many have wider responsibilities in the field of education cases, has caused unclear assignments in areas such and social welfare. In Croatia, counties4 fulfill a compa- as urban planning, environmental protection, educa- rable function at the regional, intermediary level, and tion and health, and social assistance. at the same time represent units of self-government. Croatian county centers and large cities are subject Bosnia and Herzegovina. BiH consists of the State to specific functional assignments and also provide Institutions, two entities, the Federation of Bosnia education, healthcare, physical planning, and local and Herzegovina (FBiH), and Republika Srpska; and development planning. Key functional assignments the Brcko District. The Brcko District is a neutral, self- per country are summarized in the overview Table 1. governing administrative unit, under the sovereignty A brief summary per country follows below. of BiH, formally part of both entities. The structure of local self-government is governed separately by both Albania. Albania’s two-tier organizational structure entities and differs significantly. The Federation of comprises 12 counties (quarks), 65 municipalities, and BiH has a two tier sub national government structure 308 communes. Functional assignments are stipu- comprised of (i) cantons; and (ii) cities and munici- lated in the Law on Organization and Functioning palities. There are 10 cantons, 78 municipalities, of Local Government, enacted in 2000. However, and 3 cities in the FBiH. The ten highly autonomous the Law assigns only generic responsibilities to cantons have full-fledged legislative and executive sub-national governments within the framework structures; the canton heads are elected by a majority of exclusive, shared, and delegated functions. of the cantonal Assembly from candidates nominated Exclusive functions enable sub-national govern- by elected legislators. The main organizational units ments exercising full administrative, service and of cantons are ministries. Each canton has its own leg- regulatory authority; shared functions are assigned islation, regulation, decrees, administrative decisions to both sub-national and central government units; that might affect the municipal level. Municipalities and delegated functions define actions subject to with more than 30,000 inhabitants can establish authorization of the central government. Generally, city status to amalgamate and coordinate policies municipalities and communes are responsible for among urban municipalities. Main responsibilities of core functions such as infrastructure and public ser- municipalities and cities include the core functions vices, cultural and recreational functions, local eco- such as local roads, water, sanitation, and solid waste nomic development and civil security. Quarks have management. There have been cases with unclear division of competence between cantons and cities or 3 In Kosovo, municipalities with a significant Serbian population municipalities, in particular in the Sarajevo Canton. have enhanced functions, such as secondary health care and education. Overlaps in functional assignments between cantons 4 The county government has a dual role and represents both a unit and municipalities are not uncommon, such as the of regional and local self-government: it is a local unit of public responsibilities for local infrastructure maintenance administration and a directly elected sub-national government. Since 2001, members of the county councils are directly elected. and social welfare. Similarly, assignments differ 6 Overview: Decentralization in the Western Balkans among cantons. Some cantons, but not all, have and enhanced competencies in accordance with taken on the responsibility of preschool and primary the Law on Municipalities. Municipalities shall school maintenance and capital investments. The RS have full and exclusive powers, as they concern the has a one tier local government structure, comprised local interest under the own competencies. Central of 61 municipalities and 2 cities. Municipalities are governments shall delegate responsibilities when responsible for basic municipal service provision, it is necessary. Certain municipalities shall have often with the assistance from the entity government. their own competencies providing for equal access The entity government is also responsible for educa- to public services in the areas of health (Mitrovicë, tion and health services. Graçanicë, Štrpce) and education (Mitrovicë); and shall have participatory right in selecting local sta- Croatia. Croatia has a two-tier sub national govern- tion police commanders.5 ment structure consisting of counties representing the regional government and municipalities and cities Macedonia. Macedonia is comprised of 84 munici- representing the local level of self-government. In palities and the City of Skopje, which consists of addition, large cities with more than 35,000 inhabit- 10 municipalities. 6 Macedonian municipalities ants are considered part of the municipalities and are responsible for a wide range of services with cities. There are 126 cities, 429 municipalities and 20 functional assignments7 as per the law on local counties. In 2001, Croatia launched a partial decen- self-government from 2002. Macedonia signed the tralization, 32 large cities and all of 20 counties have European charter for local government in 1996 and been assigned additional responsibilities. The county ratified it in 1997. In 1998, the Ministry of Local carries out matters of regional significance, particu- Government (MoLG) was established. The process larly matters related to: education, health care, physi- of decentralization was further supported by two cal and town planning, economic development, traffic important documents adopted in 1999, namely the and the transportation infrastructure. A county is at Government program and the Government strategy the same time a unit of deconcentrated local admin- for reforming the public administration. As a result istration and of local self-government. This dual role of these two initiatives, a working team within the affects the work of the counties’ bodies, which also MoLG was established in March 1999, to start the have a certain dualism. Cities and municipalities in process of decentralization. The Ohrid framework their self-governing area of competence carry out agreement in 2001 ended the ethnic clashes and matters of local importance through which the needs gave new momentum to the decentralization ini- of citizens are met directly and which are not by the tiatives of 1999. In 2005, the government adopted Constitution or law assigned to bodies of central a detailed plan for the transfer of competences and government such as community and housing plan- ning, physical planning and zoning utility services, 5 http://www.assembly-kosova.org/common/docs/ligjet/ child-care, social welfare and primary health care. 2008_03-L040_en.pdf 6 The 10 municipalities comprising the City of Skopje are: Aerodrom, Butel, Čair, Centar, Gazi Baba, Gjorče Petrov, Karpoš, Kisela Voda, Kosovo. Kosovo has a one tier government structure, Saraj, and Šuto Orizari. comprised of 38 municipalities. According to the 7 A wide range of responsibilities are listed in the provisions of Law on Municipalities enacted in 2008, only the article 22 of this law which include: urban planning and space arrangement; environmental and nature protection; local eco- capital city Pristina, shall be regulated by a sepa- nomic development; communal services; cultural development, rate Law due its population size and importance in accordance with the national program for culture; sports and recreation; social care and child protection; foundation of educa- as capital city. Municipalities are the basic unit of tion, financing and managing of primary and secondary schools local self-government. The organs of a municipal- in cooperation with the central government; organization of trans- port and food supplements for students and student housing; ity are the Municipal Assembly and the Mayor health care, managing the system of public health organizations which shall have the right to enact acts and take and primary health care; undertaking measures for the protection and rescue of citizens and material goods in the case of destruction any implementation measure within their areas of in war, natural disasters and other accidents; fire protection that competence and shall be elected for a term of four is provided by the local fire departments; and supervision over activities regarding the municipality’s responsibilities and other years. Municipalities shall exercise own, delegated matters determined by law. 7 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans resources to local authorities. The process of fiscal central and provincial scope and can be delegated to decentralization is envisaged by the law on financ- municipalities and cities for the purpose of efficient ing local government from 2004 to evolve in two execution, including voter records and inspectoral phases. The first phase started on in July 2005 with supervisions. Historically, sub-national governments the introduction of earmarked grants. In the second in Serbia have less responsibilities for education, phase, block transfers were launched. The major health care and social welfare. However their role principle of this phased approach was to project a in the delivery of social services has been increasing gradual devolution of responsibilities proportional over the last couple of years. to the demonstration of greater capacity by local governments to undertake those responsibilities, 2.2 SIZE AND STRUCTURE OF and to provide an equitable and adequate transfer of funds for an efficient and ongoing execution of MUNICIPALITIES transferred competencies. Municipalities in the SEE region are very diverse. Montenegro. Montenegro has one tier structure They vary significantly in size, income, and density, comprised only of municipalities. There are 21 among others. Differences among sub-national municipalities and their roles are defined in the Law governments arise, among others, from variations on Local Self Government which could be charac- in delineations and the structure of the sub-national terized as positioned between having only core government system. The following analysis pres- functions and extended local autonomy. Main pub- ents variations of sub-national governments within lic functions related to health, education and social and across the SEE Region by reviewing their size welfare are centralized and therefore, financed by and density. A classification of municipalities by the central government budget. Municipalities are urban and rural is only available in Macedonia, assigned with core functions and certain extended so specific trends in those two categories will be functions, excluding the provision of power, local reviewed in more detail in Chapter 4. The analy- police, basic health services, and construction and sis in this section focuses on the lowest level of maintenance of school buildings and hospitals. sub-national governments, i.e., municipalities, Although the payment system is supported by towns, and cities, to allow for comparison across well-functioning Treasury Single Account, almost the region. Intermediate government tiers, such as fifty percent of the municipalities do not keep cantons in the FBiH or quarks in Albania, are not records and do not report on budget execution considered in the following review. according to functional classification, which forms the requirement for a capacity building in order to The distribution of municipalities by population keep consistent records. size varies considerably in SEE countries. The average population size of local government units Serbia. Serbia has a one tier government structure in one tier sub-national government systems ranges comprised of 122 municipalities and 23 cities, plus between around 23,000 inhabitants in the RS in BiH to the Autonomous Province of Vojvodina. Cities are close to 59,000 in Kosovo (see Table 2). Here, the share municipalities with a population above 70,000. of really small municipalities which are below the Executive power of municipalities and cities is European population average of 5,000 is not higher held by the council, while the legislative power than 30% of the total local government units. On is held by the assembly. Cities and municipalities the other hand, countries with two-tier sub-national are regulated in the same way with close to similar governments like Albania and Croatia have a consid- functions. Municipalities and cities exercise origi- erable amount of municipalities with a population nal and delegated functions. They have complete below 5,000 inhabitants. In Croatia, this group of autonomy in exercising original functions, such as fairly small sub-national governments account for local utilities, urban planning and environmental the large majority of all municipalities (71.7%); and protection. Delegated functions fall within the in Albania, close to half of the total (44.2%). 8 Overview: Decentralization in the Western Balkans TABLE 1: OVERVIEW OF SUB-NATIONAL GOVERNMENT FUNCTIONAL ASSIGNMENTS IN SOUTH-EAST EUROPE Level of sub-national government (tier) Assigned function Comment Albania Counties (Quarks) • Social Assistance Central government responsible for curricula and hiring staff in edu- • Environmental Protection cation and health. • Education and health (shared) Municipalities and • Urban Planning communes • Social Housing • Legalization • Urbanization • Water Supply and sanitation • Maintenance and construction • Solid Waste Management • Road Infra structure • Social Care • Veterinary Services • Local Economic Development • Public Transportation Bosnia and Herzegovina Federation of Bosnia and Herzegovina Cantons • Pre-school Functional assignments and shared functions differ across can- • Primary Education tons. All cantons are responsible for wages and salaries of teachers; • Social Welfare municipalities usually cover operation and maintenance cost; but Municipalities • Water supply and sanitation some cantons maintain full responsibility for pre-school and pri- • Solid waste management mary education (including operation, maintenance, and capital • Local Roads and Public Transportation expenditure). • District Heating • Pre School • Primary Education • Basic health (ambulance services) • Birth and death certificates • Building permits and cadastre • Business Licensing • Culture, Sport and Leisure • Communal Inspections • Social Welfare • Housing Republika Srpska Municipalities • General public services There is a one tier structure in the RS. The government structure is • Defense very centralized. Although many functions are ensured under the • Public order and safety responsibilities of cities and municipalities, impact and the broad- • Economic affairs ness of these functions are limited. • Protecting the environment • Health care • Recreation, culture and religion • Education • Social protection (continued on next page) 9 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 1: OVERVIEW OF SUB-NATIONAL GOVERNMENT FUNCTIONAL ASSIGNMENTS IN SOUTH-EAST EUROPE (continued) Level of sub-national government (tier) Assigned function Comment Croatia Counties • Education There is a two tier structure in Croatia: Cities -Municipalities and • Health care Counties. Municipalities and cities represent the local level of • Physical planning and zoning self-government and counties represent the regional level of self- • Economic development government. In order to facilitate the service, some functions are • Traffic and transport infrastructure given under the responsibilities of large cities (cities with more • Planning and development of the net- than 35,000 inhabitants) and county centers different than cities work of educational, medical, social and and municipalities such as delivery roads maintenance and issuing cultural institutions bond and location permits Counties in their self-governing scope • Issuing of building and location permits of authority are responsible for the functions of regional character. and other document in relation to con- Financially strong local governments, 32 large cities and 20 counties, struction in the county area excluding have been assigned additional responsibilities to provide some of the area of the big city public functions locally such as road maintenance and issuing build- Large cities and • Cities and Municipality functions + ing and location permits. county centers • Public roads maintenance • Issuing of building and location permits Municipalities • Community and housing planning • Physical planning and zoning • Utility services • Child-care • Social welfare • Primary health care • Primary school education • Culture, physical culture, and sports • Consumer protection • Protection and improvement of the • Natural environment • Fire protection and civil defense • Local transport Kosovo Municipalities • Social protection There is a one tier structure in Kosovo. Municipalities shall exer- • General Public Services cise own, delegated and enhanced competencies in accordance • Protection Order and Public Safety with the Law on Municipalities acted in 2008. Some municipalities • Economic Relations have enhanced competencies in Secondary Health Care, University • Protecting the Environment Education, Area of Culture and Local Police Station Commanders. • Housing and Community Local governments with a large Serbian population have enhanced • Health functions, such as secondary health care, education, culture, local • Public Education police, etc. • Recreation, Culture • Religion Education Macedonia Municipalities • Health Improvement There is a one tier structure in Macedonia. Only City of Skopje, terri- • Environment Protection torial organization, comprises 10 municipalities and operates under • Education different law. • Social Protection • Development Programs • Sport • Culture • Communal Activities • LED • Urban Planning • Administration (continued on next page) 10 Overview: Decentralization in the Western Balkans TABLE 1: OVERVIEW OF SUB-NATIONAL GOVERNMENT FUNCTIONAL ASSIGNMENTS IN SOUTH-EAST EUROPE (continued) Level of sub-national government (tier) Assigned function Comment Montenegro Municipalities • General Public Services There is a one tier structure in Montenegro. Municipalities are • Public Order and Safety assigned with core functions and certain extended functions, • Economic affairs excluding the provision of power, local police, basic health ser- • Environment Protection vices, and construction and maintenance of school buildings and • Housing and community affairs hospitals. • Sports Culture and Religion • Education • Social Protection Serbia Municipalities • Housing and community Development There is a one tier structure in Serbia, plus the Autonomous Province • General Public Services of Vojvodina. Municipalities and Cities exercise original and del- • Economic Activities egated jurisdictions. They have complete autonomy in exercising • Education original jurisdictions such as utilities, urban planning and environ- • Social Protection mental protection. Delegated jurisdictions fall in within the republic • Sport and Culture and provincial scope and delegated to municipalities and cities for • Environmental the purpose of efficient execution such as voter records and inspec- • Public order and safety toral supervisions. Historically sub national governments in Serbia have less jurisdictions in the spheres of education, health care and social welfare. However their role in the delivery of social services has been increasing over the last couple of years. Despite an overall large number of small municipal- Bosnia and Herzegovina, Macedonia and Serbia, on ities, the highest population share in SEE countries the other hand, have a large number of medium size concentrates in medium and large cities. In Albania municipalities which are home to the biggest share and Croatia the majority of the population lives in of the population. In Montenegro, a large number of cities having between 10,000 and 40,000 inhabitants. medium size municipalities prevail, but the largest share (30%) of the population lives in the capital city. Figure 7: Distribution of the population across Despite the seemingly balanced population different municipalities sizes distribution, there is a spatial concentration in metropolitan areas. Figure 8 suggests that “Urban 100% Primacy”, i.e., a large urban center dominating a 90% country’s population distribution and economic % population per size range 80% 70% development, would not seem to be a major issue 60% in the SEE region. However, this perception might 50% be misleading since in most countries, capital cit- 40% 30% ies are composed of a number of municipalities 20% with different administrative entities. For example, 10% Belgrade in Serbia comprises of 10 “urban” munici- 0% palities and 7 “suburban” municipalities; and Skopje Albania BiH Croatia Kosovo Macedonia Montenegro Serbia in Macedonia consists of ten municipalities. While statistically these municipalities might be identified as independent Local Government Units, together, Less than 5000 5000–10000 10000–40000 they constitute one metropolitan area. The difficulty 40000–100000 100000–200000 More than 200000 of defining, re-defining and merging physical and 11 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 2: LOCAL GOVERNMENT UNITS IN SEE COUNTRIES AND POPULATION DISTRIBUTION Number of Average population LGUs below pop Capital city in % Average Density Population LGUs of LGUs 5000 (%) of total (people/sq.km) Albania 4,202,098 385 11,460 41.0% 14.5% NA FBiH 2,337,660 80 29,591 13.8% 18.7% 268.75 BiH RS 1,433,038 63 23,114 25.4% 15.8% 63.42 Croatia 4,290,582 576 7,751 70.5% 18.5% 97.31 Kosovo 2,236,963 38 58,867 5.3% 11.4% 290.47 Macedonia 2,529,473 85 26,967 21.3% 25.0% 85.77 Montenegro 620,029 21 29,525 14.3% 30.0% 61.69 Serbia 7,748,519 170 55,836 0.6% 21.0% 395.22 EU-27 average — — 5,580 — 7.0% 116.92 administrative boundaries is enhanced by the subur- As presented in Table 2, concentration on the capital banization of cities. For instance, in Tirana, although city is particularly visible in Montenegro, with 30% many towns and villages have been merged with of the total population residing in the capital city the capital city, some large suburbs such as Kamez Podgorica. In addition to the largest population (76,000) and Kashar (20,327) continue to be listed share, capital cities generally also generate the larg- as separate entities. In Montenegro, the capital city est share of GDP in a country. In Macedonia, Skopje Podgorica consists of two administrative units, accounts for about 25% of the total population and the sub-municipality of Golubovci and the sub- contributes to around 45.6% of the country’s GDP. municipality of Tuzi. In Serbia, the largest country in the region, Belgrade concentrates about 21% of the population. From the Overall, capital cities dominate the urban structure 7 countries in SEE included in the analysis, 4 have in the relatively small countries of the SEE region. more than 15% of their population living in the capital city. As a contrast, only 7 countries in the EU have capital cities concentrating more than 15% of Figure 8: Distribution of municipalities across their total population (Austria, Bulgaria, Cyprus, different sizes Estonia, Hungary, Latvia and Lithuania). 100% Municipalities in SEE are also very diverse in terms % of citiesin each population range 90% of population density. Variations in the group of 80% 70% largest municipalities with more than 200,000 inhab- 60% itants range from more than 1,600 inhabitants per 50% km2 (Serbia) to less than 200 per km2 in the RS (see 40% Table 3). For urban areas, large variations in density 30% 20% might be a reflection of different public policies that 10% provide intended or unintended incentives for higher 0% or lower densities. However, it is also the result of Albania BiH Croatia Kosovo Macedonia Montenegro Serbia divergences in the delineation of municipalities’ administrative boundaries as described earlier. For this reason, a comparison of densities should focus on Less than 5000 5000–10000 10000–40000 differences within and not across countries. Table 3 40000–100000 100000–200000 More than 200000 presents comparisons of municipalities with different 12 Overview: Decentralization in the Western Balkans population ranges for each of the countries. Overall, their rural or agricultural character. Within countries, smaller municipalities have lower population densi- divergences in densities are particularly striking in ties than larger cities which might be a reflection of Croatia, Kosovo and Montenegro. TABLE 3: MEAN DENSITY (INHABITANTS PER KM2) BY SIZE OF MUNICIPALITIES BiH RS FBiH Croatia Kosovo Macedonia Montenegro Serbia Average density in habitants per km2 less than 5,000 20.4 56.3 54.5 134.6 21.7 5.2 8.2 5,000–10,000 67.8 110.6 95.6 136.4 34.8 16.9 76.9 10,000–40,000 70.9 90.1 209.7 499.7 105.9 85.9 61.5 40,000–100,000 93.7 757.7 442.5 180.3 118.1 51.7 1,015.8 100,000–200,000 148.3 853.3 1,941.1 229.8 207.2 129.0 612.9 more than 200,000 182.8 1,236.3 556.2 870.1 1,662.1 13 3 Local Government Finance: Structure and Composition of Sub-National Budgets 3.1 WEIGHT AND INSTITUTIONAL STRUCTURE OF LOCAL GOVERNMENT FINANCE 3.1.1  Importance of local governments in the SEE region Overall, the importance of local governments has increased across the SEE region. After more than a decade-long reforms and based on the decentralization traditions in SEE, the overall framework of inter- governmental fiscal architecture has largely been put in place across the region. National legislation on local finances and local taxation set the basic rules of intergovernmental fiscal relations, financial management and local revenue raising. New grant allocation methods and revenue sharing techniques have been built into the fiscal planning practices at macro level. Table 4 provides a summary overview of local govern- ment finance, including sub-national government’s proportion of total public services, their main function, and fund profile. However, the economic weight and share of local government financing varies. Overall, sub-national governments’ financial weight is defined by (i) the scope of overall public sector and general govern- ment expenditures; (ii) the proportion of public expenditures delivered by sub-national governments; and (iii) the role of sub-national gov- ernments in delivering these public services in a particular country. On average, sub-national government expenditure in the SEE region accounts for around 5 percent or more of GDP, except for Albania with only 2.2 percent. Overall, government expenditure increases in line with the GDP. Countries with higher per capita GDP (Croatia, Montenegro, and Serbia) spend more than 40% of GDP on public ser- vices, which is in line with the averages. Others with lower GDP per capita (Albania, Kosovo) have narrower public services, 30% of GDP or below. There are exceptions to this rule: Bosnia and Herzegovina has a complex government structure with high public spending (49.2%) and Macedonia, with more limited public sector spending (27.5%) compared to its level of economic development. 15 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 4: MAIN FEATURES OF LOCAL GOVERNMENT FINANCE IN 2011 General government Local Local Expenditures as % Main Function (% of Funding – Own expenditures in % Expenditures of General Government overall local government source funding out of GDP* as % of GDP Expenditures1 expenditure) of total funding (%) Albania 28.5% 2.2% 8.8% General Public services (56%) 31% BiH (RS/FBiH) 49.2% 1.6/0.0%2 4.1/0.0 —/General Public services (48%) 82%/46% Croatia 42.5% 6.8% 16.6% Education (19%) 37% Kosovo 30.0% 8.4% 48.6% Education (48%) 15% Macedonia 31.2% 5.6% 20.4% Education (49%) 46% Montenegro 43.8% 3.9% 8.5% General public services (64%) 55% Serbia 45.2% 5.7% 13.7% — — Source: From IMF WEO data Note: 1 General Government Expenditures from FACE 2 Data was not available for all LGUs therefore this number might under-estimate the weight of local expenditures in the economy. On average, local governments are in charge of by devolution of basic municipal services, such around 15% of overall public expenditures. Sub- as local administration; but also economic affairs, national governments’ role in public service delivery including public transportation; housing and com- and administration can be measured by the size of munity amenities, such as water supply and solid local expenditures as share of overall public sector waste management (see also Table 1). In a smaller spending. By this measure, the level of decentraliza- group of countries (Macedonia, Kosovo, Croatia), tion would be highest in Kosovo, with more than sub-national governments have wider responsi- 45% of all public services being delivered by local bilities in public education—a costly public service government—more than three times higher than which, as a result, dominates sub-national budgets in Albania. The remaining countries have a more in that group. ‘General public services’ account for similar pattern with a share of around 12–20% public about two thirds of expenditures in Montenegro; sector spending by sub-national governments. ‘education services’ take up 49% of expenditures in Macedonia; while in Croatia, the main function, However, the funding arrangements are very education, accounts for only 19% of expenditures. diverse, with the own source revenues share vary- ing between 15–82%. Variations in the revenue Local government share, function and funding composition across the SEE region are large. Usually, source differ significantly in Albania and Kosovo. local government revenues are grouped as (i) own While the overall public sector in Albania and source revenues; (ii) shared revenues; (iii) transfers Kosovo is markedly smaller than in the other three and grants; and (iv) proceeds from borrowing (see countries, the importance of sub-national govern- Box 1). With 55 percent, sub-national governments ment varies almost by a factor four: sub-national in Montenegro rely on own sources somewhat 50% governments in Albania account for 8.8% of total more than in Croatia, where own funding accounts public expenditures, while in Kosovo, almost 49% of for 37%. With own source funding of 46%, sub- all public expenditures are handled by sub-national national governments in Macedonia lie in between governments. In Albania, ‘general public services’ the two other countries. No information is available represent the main function of sub-national govern- on the main sub-national government function in ments (56% of all sub-national expenditures), while the BiH RS. in Kosovo, education accounts for some 48%. At the same time, comparing revenue sources across all six Sub-national expenditure structure differs with countries, Albania and Kosovo are the two countries the functional assignments. The expenditure profile with the lowest share of sub-national government’ depends on the type of services devolved to local own funding, with 31% and 18% respectively. governments. Most of the countries are characterized For Kosovo, this implies that while sub-national 16 Local Government Finance: Structure and Composition of Sub-National Budgets of own-source revenues (MoF aggregate data). For Box 1: Classification of local government revenue sources a regional overview of revenue assignments, please see Table 5. A summary is provided in the following. Own revenue: Criteria for categorizing a tax as own revenue are the authority to define the tax base, the exemptions and the tax rate. If local governments have control over at least one of these Albania. The Albanian Law on Organization and elements of revenue policy, then the tax source is regarded as an Functioning of Local Governments requires that own revenue. the fiscal self-sufficiency of local governments shall Shared taxes: Four criteria define whether a local government rev- be guaranteed through diversified revenue sources enue is a shared tax: a) A sharing ratio is defined in law; b) shared which are local taxes, and fees arising from the revenue is un-conditional; c) the ratio is pre-determined and rental/sale of local governments services; grants predictable at least for a period of a year; d) the shared revenue from national governments, shared taxes and local is proportional to the revenues raised in the locality. If all criteria borrowing. However, legislation about local taxes are not met, the shared resources should be regarded as grants or transfers. They cannot be classified as own revenues. and fees is confusing and in some cases contradic- tory. Especially fee and tax revenues are not well Grants, transfers and subsidies: Cover other transfers from the national budget. They are provided for various purposes: general classified and create problems at the reporting stage. grant, conditional grants, and equalization grant, and vary in Almost 70 percent of the revenues are formed by governance (set by law/budget/admin procedure), and hence in Small Business Tax, Property Tax, Infrastructure visibility and predictability. Impact Tax and Fee on Solid Waste Collection. Bosnia and Herzegovina. In terms of revenues, indirect taxation (VAT, customs, excises, and road tariffs) government are in charge of more than a quarter of are defined and administered at the State level. all public expenditures, only a minor part of these The indirect taxation legislation prescribes that out expenditures, 15%, is funded by the sub-national of total revenues, first revenues needed for State governments own revenue efforts. Institutions are taken out. Then, Brcko District gets 3.55% of remaining funds. Based on a formula Institutional arrangements of local 3.1.2  derived from the final consumption data on VAT government revenues forms, the remaining funds are divided between the two entities, FBiH and RS. Out of the funds available Fiscal reporting systems in SEE do not always fol- for each entity, funds needed for foreign debt servic- low the standard categorization.8 In Macedonia, ing are subtracted first and the remaining funds are for example, VAT is shared by formula, but munici- distributed among the government levels within the palities account it as own revenue because the law entities according to entity laws. defines VAT under the account of own revenues. In spite of that, central government sees the VAT as an In the FBiH, 36.2% of indirect taxation revenues unconditional transfer for equalization purposes. belong to the FBiH Government; 51.48% to the In Croatia, the shared PIT can be supplemented by Cantons (based on population, area, number of a local surcharge, which then qualifies as local own students, and also developmental level of the source revenue. However, since PIT revenues are Canton); 3.9% to the Road Directorates; and 8.42% reported jointly, it cannot be separated and the full to municipalities and cities. This is not a shared tax, amount of PIT is reported as part of the shared rev- since LGUs receive this revenue based on a formula enues. In Serbia, the shared revenues are an impor- fixed by law rather than based on proportion of tant source, but are reported as an integrated part revenues actually raised in that LGU. The indirect 8 For regional comparison in this report, the collected data on rev- taxation revenues belonging to LGUs are distrib- enues were reclassified, where feasible. Taxes shared by origin uted to the LGUs with a formula which is mostly were separated from own source revenues and intergovernmental transfers. Revenues shared by a formula were merged with gen- based on weighted average population and other eral grants. Local taxes were reported as own source revenues, criteria such as area, weighted number of students even if municipalities cannot influence the tax base, the rate and, often, the administration is at the national level. in primary education and development index. Social 17 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans contributions (pension, health and unemployment to the supported area that a local government unit benefit funds) and direct taxation are administered at belongs to. However, Local government revenues are the Entity level and do not belong to the LGUs. In the quite unstable and are exposed to frequent changes Federation, corporate income tax revenue belongs in the model of income tax revenue sharing and to the Cantons. In terms of revenues for personal grants from the state budget. The surtax is also the income taxes, almost two third of revenues go to most generous source that makes up about 4% of Cantons, while the remaining goes to municipalities. total local government revenue. Other tax revenues In the Sarajevo Canton, this is further complicated are not abundant and are on average only about 0.5% by a large share of revenues going to the City of of local government revenue. Sarajevo, which reduces the LGUs’ share. Property tax is administered by the local government level Kosovo. In implementing the Law on Public and revenues belong to the LGUs, although in some Management and Accountability the Ministry of cases, this is also shared with the Cantons. In the RS, Finance and Economy has issued the Administrative 72% are allocated to the central government; 24% Instruction 2007/2 on Own Source Revenues of to municipalities and towns; and 4% to the public Municipalities, which provides for the types of company “Republika Srpska Roads”. In addition, municipal own revenues, as follows: property tax, Personal income tax is shared between the state tax on municipal services, tax on municipal admin- budget and municipal budgets in the proportion of istrative services, traffic fee and fees for violation of 75:25; share of fees of use of agricultural land, water municipal regulation. The own tax yield in Kosovo charges, use of mineral resource changes 70–50 to is not adequate to meet local public spending 30–50 percent between state and municipalities. requirements, and grants and transfers provide for Property tax, municipal administrative fees, utility an important part of the overall funding. 9 taxes, special water fees—fees for protection against water, municipal fees for the use of natural and other Macedonia. In accordance with the law on financing resources of general interest, taxes on gains from local governments, Macedonia initiated the second games on chance are considered as the own source phase of the decentralization in 2007 which actu- revenues of the municipalities. ally showed its effects directly on the revenues of municipalities mainly by the block grant transfers Croatia. The Law on Local and Regional Self- which were started to be allocated to the munici- Government Financing is the main legal framework palities. Transfers from central government have of the local government finance system in Croatia. highest share in the total revenues of local govern- Local and regional self-government units realize ment in Macedonia by around 60 percent. Own revenues through: own sources, shared taxes, and revenues has the second highest share almost by 30 grants from the state and county budget. Local percent. Property tax, share of personal income tax government units are entitled to the revenue from (PIT), non-tax revenues from communal taxes and shared taxes and fees collected within their area, at administrative taxes, capital revenues from the sale a percentage stipulated by the law. Shared taxes are of assets, share from value added tax are the main income tax and tax on real estate transactions. Grants accounts of the own revenue of the municipalities. are used as funds to support local government units with poor fiscal capacity. Local governments’ own Montenegro. The Law authorizes municipalities to sources include income from local governments’ introduce and levy local public revenue types, which own property, from county, city, town or municipal are municipalities’ own revenues: local taxes (surtax taxes, from fines, fees and charges. The income real- on personal income tax and real estate tax), fees ized through income tax is distributed in a manner and charges (local communal fees and local admin- that it distinguishes whether municipality, city or istrative fees, fee for construction land infrastruc- county finances or not decentralized functions in ture development, fee for use of municipal roads, selected public services including education, health care, social welfare and firefighting and with respect 9 http://www.osce.org/kosovo/31640 18 Local Government Finance: Structure and Composition of Sub-National Budgets environment fee), to administer it, to determine Figure 9: Grants and transfers still represent a tax rates within the limits prescribed by the law, to considerable share of sub-national revenues provide for tax reliefs and exemptions, to perform billing, collection and control of local revenues and Regional Comparision — Evolution of total revenue 2008–2011 to introduce penalty measures. In addition to own 100% revenues, Montenegrin municipalities receive funds 75% in the form of “assigned revenue”, i.e. shared taxes. 50% These are shares from the personal income tax (12%; with the exception of the Historic Capital which 25% receives 16% of revenues from personal income tax, 0% the Capital City which receives 13%), the tax on real –25% estate transfer (80%), fee for use of motor vehicles –50% (30%) and concession fees (70%). –75% Serbia. The local government affairs are funded –100% 2008 2009 2010 2011 from: own and shared revenues, transfers, proceeds Albania Federation of BiH BiH RS Croatia from borrowing and other revenues and earnings Kosovo Macedonia Montenegro Serbia set forth by the Law on Local Government Finance. The most important sources of local government current financing are: shared income tax (35% on average), other local revenues including construc- tion land development fee (26% on average), and main source of funding, which significantly limits grants and transfers (21% on average). Property tax local revenue autonomy. with an average share of only 11% has potential to be increased in the future. Total local revenues have increased in Albania, Kosovo and Macedonia, but took a hit in the other countries. Kosovo municipalities benefitted from an 3.2  LOCAL GOVERNMENT overall increase of close to 92% during 2008–2011, REVENUES driven largely by growing grants and transfers such as in Macedonia (40%). However, while Kosovo and Overall, grants, shared revenues and transfers Macedonian municipalities also enhanced their own dominate local revenue composition. Only in the RS revenue sources, own-source revenues declined over and Montenegro, own-source revenues account for the same period in Croatia, the RS and Montenegro. more than half of sub-national budgets (see Figure 9). Total local revenues shrank the most in Serbia and However, the revenue structure differs across coun- Montenegro: in 2011, Montenegro sub-national tries. In Montenegro, own source revenues make the governments collected 39% less compared to 2008. largest share, comprising of tax on property, PIT sur- The decline was less dramatic in Croatia (–18%) and charge and fees. Even if the property turnover tax is the RS (–12%). reported as shared revenue, own source revenues are the largest item. Own source revenues also dominate At the same time, the overall share of own source in the RS, where they comprise 82% of total revenues. revenues in total revenues declined across the SEE In Macedonia, various earmarked transfers from line region. In all countries, the overall share of own ministries dominate local budgets, so local revenue source revenues in municipal budgets has decreased autonomy is lower, despite the fact that shared VAT during 2008–2011 (see Table 6). For instance, in was not separable and is accounted as own revenue. Kosovo, despite an increase of 33% of own source Grants and transfers account for a particularly revenue in the period of analysis, the share of own large share of sub-national budgets in Albania and source revenue decreased by –8% due to the consid- Kosovo. In Albania, earmarked transfers are the erable increase in grants and transfers. 19 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 5: OVERVIEW OF SUB-NATIONAL GOVERNMENT REVENUE ASSIGNMENTS IN SOUTH-EAST EUROPE Level of sub-national government (tier) Sources of Revenues Comment Albania Counties (Quarks) • Small Business Tax (SBT) There is some confusion in defining exactly which Municipalities and • Property Tax (on buildings & agriculture land) are the fees and the local taxes, and in some cases communes • Infrastructure Impact Tax they are interpreted /used in different ways at local • Tax on the Immovable Property Transactions government level. • Tax on Vehicles (shared tax) • Fee on Solid Waste Collection (cleaning fee) • Fees for Occupying Public Space • Fees on Billboards Bosnia and Herzegovina Federation of Bosnia and Herzegovina Cantons • 51,48% of indirect taxation In terms of revenues, indirect taxation (VAT, cus- • corporate income tax toms, excises, and road tariffs) are defined and • personal income taxes, 65,5% of revenues go to Cantons administered at the State level by the BiH Indirect Municipalities • 8,42% of indirect taxation Taxation Authority. The indirect taxation legislation • 34,5% of personal income tax prescribes that out of total revenues, first revenues • Property tax needed for State Institutions are taken out. Then, • fees from lending and equity, Brcko District gets 3,55% of remaining funds. Based • positive foreign exhange revenues, privatisation, and on a formula derived from the final consumption • revenues from premium and guarantees data on VAT forms, the remaining funds are divided between the two entities. Republika Srpska Municipalities • 25% income tax Revenues from indirect taxes, paid to the budget of • Fee for change of use of agricultural land %70 the RS from the Unique account of Indirect Taxation • Concession fee for the use of mineral resources %70 Authority of Bosnia and Herzegovina, which, after • Property tax, separation of funds for needs of the external debt • Fines imposed in misdemeanor proceedings for offenses of the RS, are divided as follows: established by decisions of municipalities, • Budget of the RS 72.0%, • Municipal administrative fees, • Budgets of municipalities and towns 24.0% • Utility taxes • Public company “Republika Srpska Roads” 4.0%. • Special water fees – fees for protection against water, • Municipal fees for the use of natural and other resources of general interest, • Taxes on gains from games on chance Croatia Counties • Tax on inheritance and gifts The income realized through income tax is dis- • Tax on motor vehicles tributed in a manner that it distinguishes whether • Tax on boats and vessels municipality, city or county finances or not decen- • Tax on gambling machines tralized functions in selected public services Municipalities • Surtax on income tax including education, health care, social welfare and • Consumption tax fire fighting and with respect to the supported area • Tax on vacation homes that a local government unit belongs to. Tax auton- • Tax on firm or name omy of local government units is limited because • Tax on the use of public surfaces the rate and the tax base determined by the central government. However, the local unit can indepen- dently determine the rates in the range claimed by the central government. A share in the distribution of revenue from tax on real estate transactions for municipalities and cities is 60 percent and for the state 40 percent. Surtax on income was introduced more than 50% of local government units (most cit- ies) accounts for 4% of the budget. (continued on next page) 20 Local Government Finance: Structure and Composition of Sub-National Budgets TABLE 5: OVERVIEW OF SUB-NATIONAL GOVERNMENT REVENUE ASSIGNMENTS IN SOUTH-EAST EUROPE (continued) Level of sub-national government (tier) Sources of Revenues Comment Kosovo Municipalities • Property tax The structure in Kosovo allows municipalities to set • Licenses and permits tax rates. Municipalities are responsible for property • Certificates and Official documents tax valuation, preparing and issuing bills,managing • Motor Vehicle Fees the property tax information system, enforcing and • Building related permits collecting property taxes as well as for the adminis- • Regulatory charges trative appeals. • Education Copayment Macedonia Municipalities • Property tax, Self-financing activities are a source of revenue • a share of personal income tax (PIT), from the local government’s budget users (schools • non-tax revenues from communal taxes and administra- and kindergartens) like the participation of parents tive taxes, and the organization of excursions. Transfers from • capital revenues from the sale of assets, central government are mostly tax revenues for the • share from value added tax (VAT) block and earmarked grants for wages/salaries for teachers and employees in education, kindergar- tens and libraries and maintenance of the schools and kindergartens and culture buildings. Montenegro Municipalities • Real estate tax The Law authorizes municipalities to introduce • Local communal charges and levy local public revenue types mentioned • Local administrative charges above (local taxes, fees and charges), to admin- • Asset revenue ister it, to determine tax rates within the limits • Land development fee prescribed by the law, to provide for tax reliefs • Surtax on personal income tax and exemptions, to perform billing, collection • Local Roads Use Fee and control of local revenues and to introduce • Fee for environmental protection and improvement penalty measures. • Income from capital (interests, stakes and shares, etc.); • Fines imposed in misdemeanor proceedings, as well as gain confiscated in that proceedings; • Revenues from concession fee for performing communal affairs and revenues from other concession activities that a municipality concludes in compliance with law; • Revenues collected by municipal bodies, services, and organizations through their own activities; • Revenues from grants and subsidies; and • Other revenues set by the law (continued on next page) 21 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 5: OVERVIEW OF SUB-NATIONAL GOVERNMENT REVENUE ASSIGNMENTS IN SOUTH-EAST EUROPE (continued) Level of sub-national government (tier) Sources of Revenues Comment Serbia Municipalities • Property taxes, excluding taxes on the transfer of prop- The original local revenues are not sufficient to erty rights and taxes on inheritance and gift finance local public expenditures in Serbia. For this • Personal Income Tax (PIT) reason the LGF Law regulates share of the LGs in the • Local administrative fees revenues from certain taxes and fees, which belong • Local utility fees to the Republic, calling them assigned public reve- • Tourist fee nue. It is a system of joint public revenues, since the • Construction land use fee Republic shall establish the tax base and tax rate, • Construction land development fee and the revenues are shared between the LG and • Environmental protection and improvement fee the Republic at the moment of collection/payment. • Revenues from concessions for public utility purposes and revenue from other concessions the LG enters into in accordance with the law • Fines imposed in misdemeanor proceedings for offenses prescribed by the LG Council and assets confiscated in such proceedings • Income from lease or use of LG owned real estate • Revenues from sale of moveable assets used by the LG and indirect budget beneficiaries • Revenue from sale of real estate • Revenue generated from activities of LG bodies and organizations • Revenue from interest on funds from the LG budget while Croatia, Kosovo and Macedonia have a more Figure 10: Trends in total local revenues – regional diversified revenue composition which includes comparison revenue from rents and capital. Findings suggest that municipalities in Croatia and Macedonia leverage a Regional Comparision — Evolution of total revenue 2008–2011 100% larger share from fees and charges, while rents play a role only in Croatia and Kosovo. 75% 50% Sub-national revenues remain volatile, despite an 25% overall increase since the start of decentralization. 0% Local own source revenues decreased across the SEE –25% Region, in particular the more volatile revenues from asset sales, urban land development, or businesses –50% fees. This decrease was only partially compensated –75% by national transfers, since central government –100% budgets also contracted during the same period. The 2008 2009 2010 2011 decline was particularly contrasted in Croatia and Albania Federation of BiH BiH RS Croatia Kosovo Macedonia Montenegro Serbia Serbia, although the source differed: Croatian local governments suffered a decline in the shared PIT; Source: Federation of BiH and RS data from aggregated data in Serbia, general transfers from the central govern- ment decreased the most. Many local governments attempted to compensate for these vertical fiscal However, weight and composition varies across and imbalances with an increase in the local revenue within countries. In Albania and the RS, the largest raising capacity. However, only in Kosovo and share of own source revenue originates from taxes, Macedonia, municipalities achieved to raise more 22 Local Government Finance: Structure and Composition of Sub-National Budgets TABLE 6: TRENDS IN TOTAL REVENUE, OWN SOURCE REVENUE AND GRANTS AND TRANSFERS % Change in 2008–2011 period (2008–2010 for BiH RS) Total revenue Own source revenue Share of Own source revenue Grants and transfers Albania 34% –1% –7.7% 59% FBiH –16.3%* NA NA NA BiH RS –10.4%* NA NA NA Croatia –12% –18% –2.8% 44% Kosovo 92% 43% –8.0% 112% Macedonia 40% 38% –0.8% 43% Montenegro –36% –39% –2.5% –50% Serbia –48% NA NA NA Source: From aggregate data Figure 11: Trends in share of total revenue from Figure 12: Trends in own source revenues – own source revenues – regional comparison regional comparison Regional Comparision — Evolution of share of total revenue Regional Comparision — Evolution of Own Source Revenue from Own source Revenue 2008 – 2011 2008 – 2011 70% 50% 40% 60% 30% 50% 20% 40% 10% 0% 30% –10% 20% –20% –30% 10% –40% 0% –50% 2008 2009 2010 2011 2008 2009 2010 2011 Albania Croatia Kosovo Macedonia Montenegro Albania Croatia Kosovo Macedonia Montenegro own revenues from other sources. In Montenegro, In the FBiH, decline in own source revenue was local revenues decreased by 39.3% leaving two mostly due to fall of property tax revenue and fall thirds of the municipalities in deficit, measured by of revenues from dividends. the balance of expenditures on core functions and own revenues.10 Decreased own source revenues in Property taxes provided a relatively stable source Montenegro are mainly the result of a sharp decrease of revenue until the recent crisis. Property related in the collection of fees and charges. In Macedonia, revenues include the ‘classical’ taxes on land and despite there being a sharp increase in own source property, but also non-recurrent property transfer revenue coming from capital, the main reason for taxes and various urban construction land fees and an increase in own source revenue in the 2008–2010 comparable land-related taxes. In many SEE coun- period is an increased collection of fees and charges. tries, property tax are levied on the surface area, but the tax base is calculated using several coefficients 10 See Obradovic, N. Country MFR report and multipliers, which reflect the differences in 23 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 13: Composition of sub-national own Figure 14: Trends in total revenue as share of GDP source revenues in the SEE Region – regional comparison Own Source revenue by category 2011 or 2010 (BiH) Regional Comparision — Evolution of Total Revenues as share of GDP 100% 12% 90% 11% 80% 10% 70% 9% 60% 8% 50% 7% 40% 6% 30% 5% 20% 4% 10% 3% 0% 2% Albania FBiH BiH RS Croatia Kosovo Macedonia Montenegro Serbia 1% 0% 2008 2009 2010 2011 Taxes Capital Rents Other Albania FBiH** BiH RS** Croatia Fines Fees and charges Social contribution Kosovo Macedonia Montenegro Serbia Source: Date for Serbia from local consultants (Public Finance Bulletin; not from disaggregated data) and BiH from aggregate data property value without allowing too high fluctua- tion in tax base. per capita in the region account for a higher share of public spending, including local government 3.3  LOCAL GOVERNMENT expenditure. In Croatia, Montenegro, and Serbia, government expenditure accounts for more than EXPENDITURE 40 percent of GDP, in line with the EU-27 average of 49.1 (2011). Government expenditure in Albania Overall, local government expenditures in SEE and Kosovo, with a lower GDP, accounts for only 30 range between of 2–8% of GDP. At that level, percent of GDP or below. countries in the SEE region are generally in line with comparable countries in the EU, like Bulgaria and The shape of decentralization influences general Slovakia (6.7% of GDP, 2011), but below the EU27 government expenditure. Bosnia and Herzegovina average of 14%. Measured by the share of general pays a premium for the more complex institutional government expenditure to GDP, these figures structure inherited as a result of the Dayton Peace reflect the typical overall trend by which govern- Agreement in 1995: although BiH’s GDP is below the ment expenditure tends to vary in line with GDP regional average, public sector spending reaches 49.2 (see Table 8). Overall, countries with higher GDP percent. At the same time, public sector spending TABLE 7: TRENDS IN OWN SOURCE REVENUE FOR THE 2008–2011* PERIOD Taxes Fees and charges Rents Fines Capital Others Aggregate change Albania 10.4% 11.6% 0.0% 0.0% 0.0% 0.0% 11.3% Croatia –12.8% –16.4% 1.8% 0.0% –43.5% –100.0% –18.1% Kosovo 24.2% 29.2% 30.7% 303.8% 0.0% 47.5% 32.8% Macedonia 12.4% 24.1% 0.0% 0.0% 792.9% –0.4% 37.9% Montenegro 17.6% –63.5% 0.0% 0.0% 0.0% 2.8% –39.3% 24 Local Government Finance: Structure and Composition of Sub-National Budgets that has been assigned to the local level over the Figure 15: Trends in total local expenditures as last decade reflected in an increasing share of local share of GDP – regional comparison government spending over time, this trend was slightly reversed since 2008. The only exceptions Regional Comparision — Evolution of Total Expenditures as share of GDP are Macedonia, where local expenditures increased 10% from 4.3% of GDP in 2008 to 5.6% in 2011, and even 9% more so Kosovo, with an increase from 5.2% to 8.4% 8% 7% in the same period. In both countries, the increase is 6% largely the result of education spending which was 5% assigned to municipalities in the observed period. 4% The weight of local government spending has 3% decreased the sharpest in Montenegro, from 5.5% 2% of GDP in 2008 to 3.9% in 2011. 1% 0% However, per capita local government expenditure 2008 2009 2010 2011 in the Western Balkans is low. On average, local Albania FBiH** BiH RS** Croatia governments in the SEE region spend EUR288 per Kosovo Macedonia Montenegro Serbia capita, compared to EUR 2986.3 in the EU27 (see Source: BiH from aggregate data and correspondent to share of total BiH GDP. Table 8). Croatia and Montenegro are an exception: with an average of around EUR500 per capita, local governments there spend almost 40 percent more in the less complex one-tier institutional structure on local public services than the second ranking of Macedonia accounts for only 31.2 percent of country on this scale, Macedonia; or even 75 percent GDP, compared to the higher level of economic more compared to Serbia, the third ranking country. development. Local government spending is much lower in the Federation of BiH and Kosovo: here, local expendi- Except for Kosovo and Macedonia, the overall tures are less than half of the average in Montenegro. weight of local government spending has decreased In the FBiH, low spending is a result of the fact that since 2008. Although the level of economic devel- a large share of typical LGU expenditures is gener- opment influences public sector spending, it does ated in the Cantons. Hence, data for FBiH need to be not determine the level of fiscal decentralization. analyzed with caution. However, by far the lowest Although the increased number of public services level of per capita spending is observed in Albania. TABLE 8: REGIONAL COMPARISON OF LOCAL EXPENDITURE General government Local Local Mean local GDP per capita expenditures in % of Expenditures as Expenditures as % expenditures per 2011* USD GDP 2011* % of GDP 2008 of GDP 2011* capita 2011* EUR Albania 8,820 28.5% 2.5% 2.2% 48 BiH RS 9,190 49.2% NA NA 243 FBiH 9,190 191 Croatia 18,780 42.5% 7.4% 6.8% 459 Kosovo NA 30.0% 5.2% 8.4% 197 Macedonia 11,370 31.2% 4.3% 5.6% 367 Montenegro 13,700 43.8% 5.5% 3.9% 506 Serbia 11,550 45.2% 6.8%1 5.7% 2902 Source: * 2010 for BiH and Serbia, all in 2008 prices. Note: 1 Ministry of Finance of the Republic of Serbia, “Public Finance Bulletin, 2008–2011” 2 Ministry of Finance of the Republic of Serbia, “Public Finance Bulletin, 2008–2011” 25 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans On average, Albanian municipalities only spend 10 Figure 16: Composition of sub-national percent of the per capita average in Montenegrin expenditure by function municipalities. This is also a result of the role of quarks in Albania’s three-tier sub-national govern- Percentage Expenditures — Functions 2011 all except BiH (2010) ment system and reflects the overall low level of 100% local expenditure as share of GDP. 90% 80% General public services typically dominate local 70% expenditure. This category, which generally covers 60% 50% services of elected organs and the local administra- 40% tion, accounts for more than half of sub-national 30% expenditure in Albania and Montenegro, or close to 20% 50% in the FBiH (see Figure 16). However, accounting 10% 0% practices may distort these numbers. There is a usual Albania FBiH Croatia Kosovo Macedonia Montenegro practice to report the minor, unspecified expenditures on the largest expenditure item. In municipalities with limited local tasks, the general public services, General public services Defense Public order as the main function, also attracts other costs. Environment Housing Health Economic a airs Education Social protection Education expenses account for the largest share Recreation culture and religion Other in Kosovo and Macedonia. Apart of these two Source: Data on functional expenditures was not available for BiH RS and Serbia countries, municipalities in SEE are generally only responsible for operation and maintenance of school buildings, with small budget allocation ranging from 6.6% in Albania and 3.6% in FBiH, to only 0.7% in highest shares in Albania (26.8%) and Croatia Montenegro. However, full responsibility for pri- (15.2%). Environmental protection services, such mary and secondary education, including teachers’ as municipal solid waste management and sewage salaries and wages, makes up the lion share of local treatment are less visible among the local govern- spending in Macedonia (48.7%) and Kosovo (42.2%); ment expenditures, because they are either under- and still accounts for 19.2% in Croatia. developed or reported by public utilities responsible for communal service provision. Within this limita- Housing, communal services and economic affairs tions, environmental services account for the highest account for the largest share of urban services. share in Croatia (3.9%) and Albania (3.2%). Typically, local governments are responsible for housing development and maintenance, water Generally, multi-tier government structures reflect supply and public street lighting, accounted under in the expenditure composition. In Albania, the ‘housing and communal services’. This category upper level government, the qark spends the most on constitutes a significant part of local budgets in economic affairs, which includes public transporta- Macedonia (19.6%), Croatia (17.2%), the Federation tion services. In Croatia, the counties (zupanija) are of BiH (17.7%), and Montenegro (14.4). Public primarily responsible for the merit based services, transportation is the largest single item accounted such as education (34% of their budget expendi- for in ‘economic affairs’. This sub-sector covers tures), despite of the fact that salaries are paid by road building and maintenance, grants provided to the central budget. Here health care services are the operation of roads, railway and other public trans- second largest expenditure item (14%). Although portation systems. If local governments are directly most of the Cantons in the Federation of BiH also involved in tourism, restaurant and hotel businesses, are responsible for health and education services, then they are also accounted here. Economic affairs MFR collected sub-national government data only are present in all local budgets, but they have the at the level of municipalities. 26 Local Government Finance: Structure and Composition of Sub-National Budgets 3.4  GRANTS, TRANSFERS AND Figure 17: Real Growth Rate Index of Grants and INTER-GOVERNMENTAL FISCAL Transfers – Regional comparison RELATIONS Regional Comparision — Evolution of Grants and Transfers 2008 – 2011 (2010 for BiH RS) Grants and transfers form an important part of 105% sub-national budgets. However, in addition to 80% their importance for sub-national revenues, central- 55% local fiscal arrangements are equally important to 30% coordinate fiscal decision making across levels of 5% government. While a number of arguments may be –20% identified in favor of increased sub-national gov- –45% ernment fiscal autonomy, strong arguments exist –70% to maintain central government’s presence and –95% influence on sub-national fiscal matters, including –120% public services provision. Central governments are 2008 2009 2010 2011 in charge of national policies on equity and effi- Albania Croatia Kosovo Macedonia Montenegro ciency of service delivery. While revenue raising may be better organized at the central level to ben- efit from economies of scale in tax administration; expenditure responsibilities may be decentralized to a considerable extent for more efficient service little importance of grants and transfers in Croatia delivery. As a result, ‘vertical fiscal gap’, or ‘grant and Montenegro should be seen in the context of the dependency’, occurs and is common in practically overall funding profile: as the only two countries all countries. However, the size and the nature of the in the SEE region, Croatia and Montenegro benefit ‘dependency’, varies significantly, including in the significantly from shared taxes as funding source. In SEE region as reflected in the MFR results. Croatia, shared taxes are the most important fund- ing source, covering about 50% of expenditures. Grants and transfers account for more than Although shared taxes play a much less important 50% in Albania, Kosovo and Macedonia. Apart role in Montenegro, they still cover 12–13% of the from Montenegro, the share of grants in total has sub-national government funding. In both countries, increased in all countries. In Albania and Kosovo, the the revenues from shared taxes remained relatively grant share was increased significantly over the con- stable during the period 2008–2011. sidered period, reaching 69% and 85% respectively in 2011, while Macedonia saw a much more moderate However, the type of grants defines dependency development. In Montenegro, on the other hand, beyond their share in local budgets. To fully grants and transfers from the central government understand and assess the inter-governmental fiscal decreased by more than 10 percent from 26.4 in 2008 system, it is important to understand the nature of to 15.4 in 2010; and only recovered again somewhat grants and transfers, in addition to their size. In the in 2011, stabilizing at 20.8 percent. SEE region, some important aspects of the level of fiscal autonomy may be highlighted. For example, Shared taxes reduce grant dependency in Croatia while sub-national governments’ expenditure in and Montenegro. In Croatia, grants covered some Albania are funded to a large extent by central 15% of sub-national government funding in 2011, up transfers (69% in 2011), the funds are almost fully from 9% in 2008. In Montenegro, on the other hand, provided as unconditional grants which leave room around 21% of expenditures were grant-funded in for spending decisions at the sub-national govern- 2011, which represented a significant decrease com- ments’ discretion. On the other hand, while grant pared to 2007 when grants covered 35%. The relative dependency per the share of local government 27 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 9: REAL GROWTH RATE INDEX OF SOURCES OF LOCAL REVENUES – REGIONAL COMPARISON Albania Own source Shared taxes Grants and Transfers Borrowing 2008 41.89% 0.00% 58.11% 0.00% 2009 41.76% 0.00% 58.24% 0.00% 2010 34.22% 0.00% 65.78% 0.00% 2011 30.77% 0.00% 69.02% 0.20% Kosovo Own source Shared taxes Grants and Transfers Borrowing 2008 21.18% 0.00% 78.82% 0.00% 2009 19.39% 0.00% 80.61% 0.00% 2010 15.44% 0.00% 84.56% 0.00% 2011 15.32% 0.00% 84.68% 0.00% Croatia Own source Shared taxes Grants and Transfers Borrowing 2008 39.65% 51.16% 9.19% 0.00% 2009 36.11% 55.13% 8.75% 0.00% 2010 36.92% 54.63% 8.46% 0.00% 2011 36.84% 48.09% 15.06% 0.00% Macedonia Own source Shared taxes Grants and Transfers Borrowing 2008 48.07% 0.91% 48.68% 0.13% 2009 41.27% 0.89% 56.11% 0.00% 2010 45.53% 1.07% 52.85% 0.15% 2011 46.32% 0.82% 50.69% 1.77% Montenegro Own source Shared taxes Grants and Transfers Borrowing 2007 49.14% 11.90% 35.45% 3.51% 2008 57.76% 10.83% 26.44% 4.97% 2009 50.54% 9.30% 33.91% 6.25% 2010 65.55% 9.71% 15.44% 9.30% 2011 55.25% 13.77% 20.79% 10.19% financing seems much lower in Croatia and 3.5  SUB-NATIONAL BORROWING Montenegro, the grant ‘picture’ is also much more AND DEBT complex, reflecting the larger and different role for sub-national governments in these two countries. Ear-marked grants are used to ensure standards Sub-national borrowing is strictly regulated across and quality of service delivery in core public sector the SEE region. All countries have borrowing laws areas like education and health, but leave no or little which control local government debt and limit sub- financial discretion to sub-national governments. national borrowing by revenue and outstanding In Croatia, e.g., central governments increased debt. In Albania for example, administration regu- grant funding to support primary and secondary lations severely constraints sub-national borrowing education delivered by sub-national governments as. All of the countries’ sub national government to compensate for shrinking revenues from income needs consent or approval of the central government taxes. For the same reasons, expenditure needs and in order to take a loan, guarantee or issue a bond. revenue base equalization is an important aspect According to the Law of Local Self-Government in of the general grant system in these two countries, Montenegro and Serbia, municipalities are allowed and equally in Macedonia, where municipalities to issue bonds. In Montenegro, municipalities may also are in charge of education. For further details, also accrue debt in a way that the total payments of please refer to Table 10 on the grants systems in the principal and interest, payments under a leasing con- countries. tract, repayment of obligations for prior period, and 28 Local Government Finance: Structure and Composition of Sub-National Budgets TABLE 10: ARRANGEMENTS ON GRANTS AND TRANSFERS IN SEE Level of sub-national government (tier) Arrangements on Grants and Transfers Albania Counties (Quarks) Unconditional transfers are funds from the central government to local governments based on the ratio of exclu- Municipalities and sive and shared functions performed by the local governments and for the purpose of achieving equalization of communes resources among local governments. Each local government have full discretion in deciding how to use the trans- fer. The unconditional transfer was introduced in 2001, and starting from 2002 the transfer is distributed based on a predefined formula. The unconditional transfer suffers the unpredictability. The overall level of the transfer is subject to annual budget negotiations, which implies that the level and the formula for distributing the transfer may change from year to year. Bosnia and Herzegovina Federation of Bosnia and Herzegovina Cantons Total grants are almost half of LGU revenues if we include indirect taxation revenues, which technically belong to Municipalities general grant category, since allocated by formula and not based on origin of revenues. Significant funds in LGU budget come from transfers from higher government levels which are not based on reallocated Indirect Taxation Authority (ITA) tax revenues (18% of total LGU revenues are these transfers), which are practically all earmarked, mostly for capital expenses. This significant part of the municipal budget could potentially depend on ad hoc decision of higher-government levels The extent to which own non-delegated functions are financed by own revenues cannot be established. However, the LGUs also sometimes use own revenues for functions which should be delegated from higher levels. For example, due to unclear competence division in the area of local infrastructure maintenance, municipalities allo- cate resources to local community priorities even though competence is shared with Cantons. Note that Cantons are not regarded as local governments and were not subject of this analysis. Republika Srpska Municipalities As noted above, significant funds in LGU budget come from transfers from higher government levels which are not based on reallocated tax revenues (66,71 of total LGU revenues), which are practically all earmarked, mostly for capital expenses. This significant part of the municipal budget could potentially depend on ad hoc decision of higher-government levels. The current grants are the least predictable, least reliable and the most arbitrary of all municipal revenues. The allocation of income tax is an important part of general grants to municipalities, with 25% of the collected revenue in the municipality being returned as a general grant. The sharing of information by central government on collected tax revenues may be improved. Croatia Counties There are two kind of grants allocated to LGU: i) current and capital grant from central government budget and Large cities and ii) general grants based on income tax revenue sharing for funding local government units with low fiscal capacity. county centers Grants from central government to local units are relatively small and account for only about 15% of total local government revenue. The amount of the grants varies significantly throughout the years. Counties do not have Municipalities stable sources of revenue since they largely depend on funding from shared taxes (mainly income tax) and grants from the central government budget. Due to the decrease in revenue from income tax, which is a major source of revenue for the county funding decentralized function of primary and secondary education, the state has signifi- cantly increased current grants. In Croatia, there are six types of current general grants of which the Ministry of Finance currently are using four of these. current general grants for municipalities, cities and counties in Croatia are: 1 grants for decentralized functions 2 current grants of the Ministry of Finance to counties 3 current grants of the Ministry of Finance to the cities PPDS first and second groups, 4 current grants to other ministries and institutions of the central government, 5 budgetary reserves and compensation for damage caused by natural disasters and 6 current general grant to replace the profit tax. The framework on allocation of general grants may require further clarity and administrative simplification (continued on next page) 29 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 10: ARRANGEMENTS ON GRANTS AND TRANSFERS IN SEE (continued) Level of sub-national government (tier) Arrangements on Grants and Transfers Macedonia Municipalities The law on financing local government identifies the following channels of transfers from the central government: 1. Value added tax (VAT) revenues (total fund equal to 3 % of the VAT collections in the previous fiscal year). This unconditional grant is distributed by a formula which states that at least 50% of the grant will be distributed accord- ing to population and not more than 50% according to other criteria. The central government sees the VAT as an unconditional transfer for equalization purposes even though it is defined in law as an own revenue for local governments. 2. Personal income tax (PIT) revenues are an unconditional grant distributed on an origin basis. The total pool is calculated as 3% of the PIT from salaries allocated to the local government where the employee resides and 100 % PIT collected from artisan activities. 3. Earmarked transfers are allocated for operational costs in the areas of education, culture and social policy. The appropriate ministries and agencies monitor the use of these earmarked funds. 4. Capital transfers are distributed in accordance with programs specified by the government. 5. Block transfers are the same as the earmarked grants plus the wages and salaries. The appropriate ministries and agencies are responsible for defining the methodology and criteria to be used in this transfer formula. Ministries define the formula by stating the scope of the transfer and the way of transfer to the specific municipalities. 6. Funds received for delegated competencies from the central to the local governments. Funds received for del- egated competencies from the central to the local governments. In this case, the amount of funds is determined by way of a contract signed between the mayor of the local government and the appropriate ministry responsible for the competency. Montenegro Municipalities In 2012, 13 municipalities are receiving funding transfers from the central government in the form of Equalization grants to meet their expenditure responsibilities. The central government makes significant grants to these munici- palities to make up their revenue shortfall resulting from vertical fiscal imbalance. These grants to the municipalities take two forms: • Equalization grants; and • Specific purpose payments or tied “conditional grants” to provide for financing municipal IPA (an Instrument for Pre-Accession Assistance) projects in 2011 and 2012. Non-earmarking of grants – The Law on Local Government Finance establishes that all transfers to municipalities in the form of “assigned revenue” and equalization payments be unconditional, meaning that there are no con- straints imposed on municipalities by the central government in deciding how to spend those transfers Municipalities are entitled to use “conditional grants” from the national budget to finance priority investment projects that are of special interest for one or several municipalities and for co-financing of projects mainly funded by donations. Serbia Municipalities The Law on Local Government Financing (LGF) regulates share of the LGs in the revenues from certain taxes and fees, which belong to the Republic. The revenues that are fully or partially assigned to the LGU are: 1. Personal income tax: 80% of PIT paid in the LGUs; 70% of PIT paid in the Capital city; and 100% of PIT from agricul- ture and forestry, self-employment, real estate, lease of moveable property, personal insurance; 2. Tax on transfer of property rights and Inheritance and gift tax (100%) 3. Assigned Fee (100%) for: motor vehicles, environmental pollution, use of mineral resources, extracted material from rivers, use of forests, water use, tourist fee etc. 4. In order to establish vertical and horizontal fiscal balance and to introduce solidarity with the economically weak, underdeveloped and small LGUs, the LGF Law regulates the obligation of the Republic to assign a part of its revenues to LGUs through the following transfers: Unconditional transfer for all LGUs should amount to 1.7% of the GDP and is allocated to: Equalization transfer belongs to each LGU which revenues from shared taxes per capita is less than 90% of average shared revenues per capita for all municipalities (cities are not included). It is calculated based on the number of pop- ulation and average shared revenues Compensation transfer is used to compensate for the revenues that are lost due to changes of the tax legislation, which is not compensated for by other revenues General transfer is the difference between the total unconditional transfer and the amount of the equalization and compensation transfer. It is shared between all LGUs based on: number of population; territory/area; number of classes in and number of buildings of elementary and secondary schools; number of children included in the child care program, number of buildings used for child care program and ration for the level of development of each LGU. Solidarity transfer is the amount of unconditional transfer calculated for the capital city, which is not paid to the City than is allocated to LGUs Conditional transfer is allocated by the Republic to LGUs for performing particular original and delegated functions 30 Local Government Finance: Structure and Composition of Sub-National Budgets any other obligations that have the character of debt TABLE 11: SUB-NATIONAL BORROWING may not exceed 10% of the realized current income % change in in the year preceding the year of borrowing.11 In % borrowing borrowing to total between % borrowing Croatia, counties are allowed to sustain guarantees revenue 2008–2011 to GDP – 2011 to municipalities to take a loan with the consent of Albania 0.21% NA 0.0071% the government. In Macedonia, municipalities can Croatia 0.00% 0.00% 0.0000% borrow with approval from the central government, Kosovo 0.00% 0.00% 0.0000% but the borrowing capacity is limited by debt service Macedonia 1.98% 1836.25% 0.1049% (30% of operational revenues) and debt stock (100% Montenegro 10.19% 30.06% 0.7249% operational revenues). Similar limits apply for BiH, Serbia* 9.80% 346% 0.6752% although there are different debt ceilings for munici- palities in the two entities. In the Federation, the debt ceiling is 10%, including loans and guarantees; while in the RS, the annual debt ceiling is 18% of loan revenues are significantly higher compared to current revenues from the previous fiscal year with the national average. a ceiling for guarantees of 30%. However, any debt service exceeding 10% of current revenues requires Imbalances in sub-national debt may become a approval from the RS Government. cause of future fiscal stress. In Albania, mostly larger municipalities and local governments in less Overall, sub-national borrowing remains a minor developed regions borrow. In Montenegro, some of source of revenue in SEE. Total accumulated local the poorer municipalities are more actively acquir- government debt is not significant. Sub-national ing loan funds, despite the heavily regulated local government debt is not significant, at less than 1% of government borrowing procedures. In 2011, bor- GDP, even in countries with more actively borrow- rowing was concentrated in Montenegro’s Northern ing sub-national governments, such as Montenegro Region (41%) and Central Region (33.5%), outside of and Serbia. However, the role of municipal credit Podgorica, where weaker municipalities were cover- would be expected to increase with greater financing ing revenue gaps by borrowing for operating expen- needs for local infrastructure investments and more ditures. In Macedonia, local government borrowing restrictive central government budgets. In addition, is only 1.8% of total municipal revenues, but two financial reports do not always include guarantees times higher in Skopje. The City of Skopje and the issued by local governments. In Croatia, for example, ten municipalities comprising the city have become issued guarantees to public utility companies equal the most active borrowers in the country: their share the total amount of total local government debt. of local borrowing accounts for 84% of total local government debt in Macedonia. In Croatia, borrow- However, larger cities and financially weak munici- ing accounts for 4.9% of local revenues, but generally palities seem to accumulate growing amounts only the cities qualify to borrow. Hence, the majority of debt. Measured by the share of long term local of debt is concentrated in the financially stronger government borrowing as percentage of local local governments. In Serbia, Belgrade is the largest revenues, findings suggest two trends: In Albania sub-national borrower: 14% of local revenues gener- and Montenegro, borrowing has become a special ate from debt; and in 2010, 73% of total sub-national supplementary funding source for weaker munici- borrowing was concentrated in the capital city. palities, while capital cities do not make large use of debt financing as a revenue source. In Macedonia Debt concentration in weaker municipalities is of and Croatia, however, mostly large cities that can particular concern. The present annual loan repay- afford to borrow make increasing use of loans. Here, ment (principal and interest) burden is the highest in Montenegro and the RS which have accumulated significant local debt. Annual debt repayment in 11 Excluding municipal borrowing “intended for the implementation of the Rehabilitation Plan for Overcoming Financial Difficulties.” percentage of the total current expenditures is 31 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans extremely high in Montenegro (53.6%) and the RS the Ministry of Finance. Although other countries in BiH (26.3%). In Montenegro, local debt is concen- accumulated lower sub-national government debt trated in the coastal cities and the local governments with more reasonable annual debt repayment rates, in the Northern region. These municipalities cannot the trend may be a concern in some of the countries cope with the high debt repayment burden, so a in light of overall growing public debt in the SEE special financial recovery plan had to be signed with region. 32 4 2 Key Trends and Findings T his section highlights a number of key observations on varia- tions across municipalities and discusses selected findings from the analysis of disaggregated local finance data. The review focuses on local government’s role in and contributions to public ser- vice delivery, but also includes observations on the fiscal structure and trends. The purpose of the section is to bring these findings into the forefront of policy considerations in the Western Balkan countries. A full-fledged analysis and assessment of these observations lies outside the confines of this report. However, the findings outlined in the fol- lowing provide insights and suggestions for further analysis based on the regional, disaggregated data made available by the MFR. 4.1  VARIATION AND DISPARITIES ACROSS TYPES OF MUNICIPALITIES 4.1.1  Population size Grant dependency and own-source revenue potential differ across municipalities within countries. The composition of local revenues differs across but also within the SEE countries. Among the studied countries, municipalities in the RS and in Croatia stand out for rely- ing mainly on own source revenues, while municipalities in Kosovo and Albania are highly dependent on grants and transfers. However, a closer review reveals significant variations within countries. For example, small and medium size municipalities in Macedonia receive a high share of their revenues from grants and transfers, while larger municipalities rely mostly on own-source and shared revenues. In Montenegro on the other hand, municipalities appear to have a more balanced mix of own-source revenues, grants and transfers, and bor- rowing, although with significant differences across municipalities of different size. For example, we observe ‘forced borrowing’ to cover deficits in smaller municipalities, while own-source revenues dominate the large ones (see Figure 18). However, the analysis needs to take into account the differences in reporting revenues among countries, 33 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans as Montenegro was one of the most detailed and Kosovo, we find a high concentration of own- complete sets of data. A comparison of variations source revenues in the larger centers, this is not of total revenue sources by city size across all MFR a consistent pattern across the SEE Region. In countries can be found in Annex 1. Croatia, municipalities with more than 200,000 inhabitants represent around 45% of the total own Local government own source revenues vary hugely source revenues of local governments, while they within the countries. Factors such as the local eco- are home to only around 18% of the population. nomic base and property values, among others, Similar results are found in Kosovo. In Albania, determine the room for collecting property related Macedonia and Montenegro, on the other hand, taxes or business fees. For instance, property values medium size municipalities concentrate most of tend to be higher in larger municipalities when com- the countries’ own-source revenues—despite the pared to small municipalities. As a result, own source lower per capita own-source revenues in those revenue potential differs across local governments. smaller municipalities (see Figure 20; for per capita In Albania, for example, it is possible to see how revenues, see Annex 2). This is mainly the result of the share of own source revenues to total revenues the large number of medium size municipalities in increases with the city size, while the share of grants this latter group of countries. and transfers decreases (see Figure 19). The same is true for the RS and Kosovo. Figure 19 also presents Smaller municipalities seem to underutilize fees the mean tax revenue per capita across different city and charges in Macedonia and Montenegro. While sizes in Albania, suggesting that bigger municipalities fees and charges account for a comparable share of tend to have higher per capita tax revenues compared total local revenues across municipalities of different to municipalities with a smaller population. size in most of the SEE countries, findings suggest that municipalities with a population smaller than However, own-source revenues are not neces- 10,000 have significant revenue enhancement poten- sarily concentrated in the biggest municipali- tial in Macedonia and Montenegro. Noteworthy ties. While in some countries, e.g. Croatia and is that municipalities in Montenegro appear to Figure 18: Variations in local revenues by population size of municipalities (Macedonia and Montenegro) Macedonia Montenegro 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than less than 5000 5000 –10000 10000–40000 40000–100000 100000–200000 200000 Own source Shared taxes Grants and Transfers Borrowing Other 34 Key Trends and Findings Figure 19: Variations in own source revenue (left) and mean per capita tax revenue (right) by city size in Albania 100% 4.5 90% 4.0 Mean Tax Revenue (LEK) per capita 80% 3.5 70% 60% 3.0 50% 2.5 40% 2.0 30% 1.5 20% 10% 1.0 0% 0.0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 more than 200000 200000 Own source Shared taxes Grants and Transfers Borrowing Other compensate smaller revenues from fees and charges Expenditure variations are largely caused by the with a larger tax income, while similarly sized service function concentration in geographical cen- municipalities in Macedonia have significantly ters. For example, in Albania, half of the total local higher capital revenues. public education spending is made in Tirana. In the Figure 20: Geographic concentration of total revenues by city size Croatia Macedonia 100% 4.5 90% 4.0 80% 3.5 70% 60% 3.0 50% 2.5 40% 2.0 30% 1.5 20% 10% 1.0 0% 0.0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 200000 Own source Shared taxes Grants and Transfers Borrowing Other 35 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 21: Composition of own-source revenues by size of municipalities (Macedonia and Montenegro) Montenegro Macedonia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Taxes Fees and charges Rents Fines Capital Other Federation of BiH, water services (housing and com- spend a higher share of their expenditures in educa- munity amenities), health care, recreational, sports tion compared with smaller municipalities. Findings and culture, and social protection services are con- also suggest that large municipalities concentrate centrated in Sarajevo and the cantonal main cities. By the bulk of expenditures in the education sector. In and large, findings suggest that the share of general Macedonia, on the other hand, large municipalities public service expenditure of total local expenditure spend much more on the housing and communal increases with decreasing size of municipalities, in services category compared to smaller and medium- particular in Croatia, Kosovo, and Macedonia; but less sized municipalities (Figure 24). so in Albania and Montenegro where this expendi- ture category shows a more equal distribution across While the largest share of expenditures and rev- municipalities of different sizes (Albania) or larger enues is usually concentrated in bigger municipali- variations (Montenegro). Overall, a limited degree ties, a per capita expenditure perspective provides of functional assignments at the local level results in nuances to this trend. Figure 25 presents variation on a large share of general public expenditures. per capita expenditures for General Public Services in Kosovo. Findings suggest that as city size increases, However, asymmetric functional assignments there is a reduction in the mean per capita expendi- amplify variations. For example, during the tures in this category. Furthermore, the analysis also decentralization process in Croatia, local govern- finds that municipalities with less than 5,000 inhabit- ment functions were devolved to the city level ants have much higher spending than medium and very selectively. Only the 32 financially stronger big municipalities. A similar patter can be found in municipalities were authorized to take over addi- the Federation of BiH. This might be a reflection of tional responsibilities, such as education, housing economies of scale, since smaller municipalities gen- and community amenities. An example of such erally face higher per capita fixed costs for provision disparities can be seen in the following figures. of a minimum service level. These results suggest From Figure 23, it is possible to see how Croatian that in some countries, existing expenditure assign- municipalities with more than 200,000 inhabitants ments do not fully leverage economies of scale and a 36 Key Trends and Findings Figure 22: Disparities of functional expenditure by size of municipalities in Albania and Kosovo Albania Kosovo 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 more than Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 200000 General public services Defense Public order Environment Housing Health Economic a airs Education Social protection Recreation culture and religion Other regionalization of some of the services may possibly Macedonia suggest overall lower dispersion of values support more efficient service provision. for per capita own source revenues. In terms of per capita capital expenditures, the RS, Federation of BiH, Overall, per capita variations are largest in the RS, and Montenegro present higher levels of dispersion. Croatia and Montenegro. In these three entities, variations of per capita own-source revenues, total 4.1.2  Urban and rural revenues, and grants and transfers suggest large dif- ferences across sub-national governments (see Figure For Macedonia, categorization by urban- 26). At the same time, results from Albania, Kosovo and rural allows for additional analysis of the Figure 23: Variations in education expenditures by city size in Croatia 7E+09 23% 25% Percentage of expenditure in Education Total Expenditure in Education by city 6E+09 20% 19% 19% 18% 20% 5E+09 from total expediture 15% 4E+09 15% range LEK 3E+09 10% 2E+09 5% 1E+09 0 0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Total Share on education 37 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 24: Variations in housing expenditures by city size in Macedonia 1.4E+10 45% Percentage of expenditure in Housing Total Expenditure in Housing by city 1.2E+10 42% 40% 35% 1E+10 from total expediture 30% 8E+09 25% range LEK 18% 6E+09 14% 16% 16% 14% 20% 15% 4E+09 10% 2E+09 5% 0 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Total Share on Housing per km2) is much higher than in rural areas (64 Figure 25: Variations on per capita expenditures in inhabitants per km2). General Public Services within municipality ranges – Kosovo Overall, urban areas generate a higher share of own-source revenues. As expected, there are consid- 400 erable variations between urban and rural munici- palities in the composition of total revenues, but 300 also the share of own source revenues. In addition to the higher share of own-source revenues, urban 200 municipalities also generate a higher percentage of own-source revenues from taxes, fees and charges. 100 0 More surprisingly, capital revenues are signifi- cantly higher in rural areas. However, this might less than 5000 5000 to –10000 10000–40000 40000–100000 100000–200000 more than 200000 be a reflection of the revenue enhancement potential in rural municipalities, which have not yet fully utilized their capacity to collect taxes, fees and charges. In addition, starting in 2009, Macedonian municipalities gained control over the land trans- actions which resulted in numerous sales of assets municipalities. Among the countries included and increase in own-source revenues. Therefore, it in the MFR, Macedonia is the only that contains is possible that capital sales in rural areas have a data on whether the sub-national government higher effect on their total revenues when compared units are considered urban or rural. From the to urban areas, which might explain the difference data, Macedonia has 41 rural municipalities and observed in Figure 27. 44 urban municipalities. Urban municipalities have a minimum of 4,545 inhabitants while rural Functional and capital expenditures patterns also municipalities report a minimum of 1,322 inhab- vary between urban and rural municipalities. For itants suggesting that there is some overlap in example, urban municipalities spend considerably terms of population numbers. However the mean more on housing, recreation and culture than their population density in urban areas (823 inhabitants rural peers. However, rural municipalities have 38 Key Trends and Findings Figure 26: Regional variations on per capita own source, total revenue, grants and transfers and total expenditures Per capita Own source revenue EUR Per capita Total revenue EUR 2,000 3,000 1,500 2,000 1,000 1,000 500 0 0 Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro Per capita Grants and Tranfers EUR Per capita expediture EUR 1,500 4,000 3,000 1,000 2,000 500 1,000 0 0 Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro higher per capita expenditure than urban munici- 4.1.3 Densities palities in the functional categories of environ- mental protection, education, health and general Urban density partly explains disparities among public services. Housing is the only expenditure local governments’ expenditures in the SEE Region. item in which urban municipalities spend more Density is relevant for the unitary cost of provision per capita compared to rural municipalities. This of selected municipal services, such as water supply might be explained by larger economies of scale in and sanitation, solid waste management, and public urban municipalities. Rural municipalities tend to transport, among others. In theory, as population have lower densities which make service provision densities decrease, it gets more expensive to provide for selected public service areas more expensive; those services. To evaluate how density affects per while smaller population size causes fixed costs to capita costs of service provision in SEE municipali- be divided among fewer people. Finally, although ties, we compare per capita functional expenditure urban areas concentrate a larger share of total capi- for selected services (environment, housing, health tal expenditure, per capita capital expenditures are and education) against the inverse of density. While slightly larger in rural municipalities. expenditures might not be the best proxy for costs, 39 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 27: Variations between urban and rural municipalities in Macedonia Total revenue 100% 1.00E+10 90% 9.00E+09 80% 8.00E+09 70% 7.00E+09 60% 6.00E+09 50% 5.00E+09 40% 4.00E+09 30% 3.00E+09 20% 2.00E+09 10% 1.00E+09 0% 0.00E+00 Rural Urban Rural Urban Taxes Rents Capital Fees and charges Fines Other Own source 100% 2.50E+10 80% 2.00E+10 60% 1.50E+10 40% 5.00E+09 20% 0% 0.00E+00 Rural Urban Rural Urban Own source Shared taxes Grants and Transfers Borrowing Other Functional expenditure 100% 2.50E+10 Total expenditure by category (MKD) – 2011 80% 2.00E+10 60% 1.50E+10 40% 5.00E+09 20% 0% 0.00E+00 Rural Urban Rural Urban General public services Defense Public order Environment Housing Health Economic a airs Education Social protection Recreation culture and religion 40 Key Trends and Findings Figure 28: Density versus Environmental Protection Figure 29: Density versus Education per capita per capita expenditures in Macedonia – 2011 expenditures in Macedonia – 2011 1,500 10,000 8,000 1,000 6,000 4,000 500 2,000 0 0 0 5,000 10,000 15,000 0 5,000 10,000 15,000 Density (hab/sq.km) Density (hab/sq.km) as they may include inefficiencies in services provi- conclusions can be made from capital expen- sion, they can provide a general idea of variations diture data limited to a one year period (here: within a country. 2011 or 2010, see Figure 30 below) and should be interpreted carefully, some observations Density effects can be observed predominantly on cane be noted. Overall, capital cities and other environmental protection expenditure, and less urban centers dominate capital spending across on health and education. In the case of Croatia and the SEE region, particularly in Serbia (Belgrade: Macedonia, we find that increasing density corre- 65%), Macedonia (Skopje: 71%) and Montenegro lates with a significant reduction in per capita expen- (Podgorica: 45%). Local capital expenditures are ditures for environment protection. A graphical more evenly distributed in other countries, with visualization of this relationship for Macedonia can the ratio of capital spending generally represent- be seen in Figure 28. These results provide evidence ing the proportional share of local budgets and to support the assumption that denser municipalities population in the capital cities. tend to have lower per capita cost for environmental protection expenditures. In the case of health, no sig- In Macedonia and Montenegro, large municipali- nificant relation was found between density and per ties concentrate the highest share of total capital capita expenditure in any of the countries analyzed expenditures. For example, in Macedonia, 71% in the MFR. In the case of education, a significant of total capital expenditures are concentrated in relation between per capita expenditures and density Skopje, the only municipality in the country with was found only for Croatia. Here, results suggest more than 200,000 inhabitants (see Figure 31). that as municipalities get denser, per capita costs The mean per capita capital expenditures are also of education increase. However, these results could considerably higher in Skopje when compared to reflect the fact that growing and larger municipalities smaller municipalities in Macedonia. The concen- tend to be both denser and have more responsibili- tration of mean per capita concentration can be ties for the provision of higher levels of education. seen more clearly in Figure 31, comparing total capital expenditures, mean capital expenditures, 4.2  CONCENTRATION IN CAPITAL and mean per capita capital expenditure per size range of municipalities. INVESTMENT EXPENDITURE In Albania, BiH and Croatia, mid-size municipali- Capital expenditures are concentrated in cities ties account for the highest share of total capital and larger municipalities. Although no major expenditure. This concentration is the result of the 41 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 30: Regional variations on capital expenditures by city size and total share % capital expenditures to total expenditures 100% 35.8% 40% 34.0% % Capital expenditures by city size 90% 35% 80% 29.4% 29.4% 25.9% 30% 70% 60% 22.1% 25% 50% 17.4% 20% 40% 15% 30% 10% 20% 10% 5% 0.5% 0 0 Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro Serbia Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 % Capital expenditures by city size large number of smaller and medium size munici- Capital expenditures constitute a major part of palities in these countries. However, despite the local expenditure in the SEE region. Municipalities similar share of total capital expenditure in these spend more than 25% of their available funds countries, the share of capital expenditure to total on capital investments (see Figure 33), with the local expenditure varies significantly: from only 0.5% exception of Croatia (17%).Overall, local capital in Macedonia to over 35% in Kosovo. expenditures per capita vary between EUR48 in Albania and EUR459 in Croatia, or 2.2% and 6.8% However, detailed analysis reveals large variations percent of GDP. Overall, variations in local capital between total and per capita capital expenditures. expenditure levels are in line with the differences For Albania, a similar analysis but with very dif- in economic development. However, Kosovo is ferent results can be seen in Figure 32. In this case, a noteworthy exception: per capita local capital the biggest share of total capital expenditures is expenditure is above the average at 8.4% percent. concentrated in municipalities with 10,000–40,000 In addition, it is important to highlight that the inhabitants. However, the largest mean capital data used for this analysis might underestimate expenditure can be found in municipalities rang- the total scale of local investments, since some are ing with more than 200,000 inhabitants (Durres executed locally but financed by the central gov- and Tirana). The per capita capital expenditures ernment. This latter type does not appear on local analysis reveals a more or less homogenous national governments’ books. scenario, although Durres and Tirana—the two big- gest cities in Albania—have the highest per capita However, functional assignments and funding capital expenditure. In Kosovo, large municipalities mechanisms have a large influence. Findings from concentrate the highest share of capital expenditure the MFR sample suggest that local governments in and present the highest average of capital expen- more decentralized countries spend more on capital diture. However, small municipalities have much expenditures than their peers in less decentralized higher per capita capital expenditures. countries. For example, the share of local capital expenditures in percentage of GDP is very similar Generally, the scope of local capital expenditures in Macedonia, Serbia and Croatia, despite the dif- varies with the level of economic development. ferent income levels (see Table 12). However, this 42 Key Trends and Findings or whether countries have unitary or federal orga- Figure 31: Macedonia: variations on capital nizational structure.12 expenditures, Skopje dominating Local governments only generate a fraction of Macedonia — Total capital expenditures by city range local capital investment funds from own sources. Capital Expenditures (Total -MKD) 200,000,000 Only a small share of local investments is financed 150,000,000 from own capital revenues. Local capital revenues, were they were reported separately (RS, Croatia and 100,000,000 Macedonia), constitute between 1.6 and 17.9 percent 50,000,000 of local own revenues. Macedonia is an exception, with own source capital revenues representing 0 around 15% of the local budget in 2011. The latter More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 can be explained by the transfer of land ownership to municipalities which started in 2009 and allowed local governments in Macedonia to be in charge of the transactions of their assets.13 Predominantly, Macedonia — Mean capital expenditures by city range capital investments are financed by grants from the central budget, extra-budgetary funds, funding from Capital Expenditures (Total -MKD) 200,000,000 international development programs, or municipal 150,000,000 borrowing. Municipalities also use current budget 100,000,000 surpluses to finance capital investments. Public companies, utilities or special asset management 50,000,000 units are also active in financing local capital invest- 0 ment projects. Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 4.3  THE ROLE OF CAPITAL CITIES Capital cities dominate total sub-national expen- ditures. In the countries of former Yugoslavia, the Macedonia — Mean per capita capital expenditures by city range member republics’ large urban centers usually 400 became the new capital cities. These larger cities Capital Expenditures (Total -MKD) 300 have often inherited extended service networks with better quality as a result of the capital invest- 200 ments and development programs concentration 100 under the socialist regime. Spatial concentration of 0 local expenditures in the capital cities is especially prevalent Albania (Tirana), Bosnia and Herzegovina More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 (Sarajevo) and Serbia (Belgrade), where the share of local expenditures significantly exceeds the popula- tion share living in the capital. In addition, in all of the SEE countries, except Albania and Croatia, there is a disproportionate concentration of capital expen- ditures in capital cities as can be seen in Figure  33. This might be explained partly by the fact that most contradicts with findings from a recent global analy- 12 Frank, Jonas and Martinez-Vazquez, Jorge (eds.). 2013. “Decentral- sis which suggests that the level of public invest- ization and Infrastructure: From Gaps to Solutions”. Georgia State ment is not directly related to the overall degree of a University and World Bank, forthcoming. country’s level of centralization or decentralization; 13 This was previously under the National Government’s control. 43 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 32: Albania and Kosovo: variations on capital expenditures, three different stories Albania — Total capital expenditures by city range Kosovo — Total capital expenditures by city range 4,000,000 50,000,000 Capital Expenditures (Total – EUR) Capital Expenditures (Total – LEK) 3,500,000 40,000,000 3,000,000 2,500,000 30,000,000 2,000,000 1,500,000 20,000,000 1,000,000 10,000,000 500,000 0 0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 Albania — Mean capital expenditures by city range Kosovo — Mean capital expenditures by city range 1,200,000 25,000,000 Capital Expenditures (Total – EUR) Capital Expenditures (Total – LEK) 1,000,000 20,000,000 800,000 15,000,000 600,000 10,000,000 400,000 200,000 5,000,000 0 0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Albania — Mean per capita capital expenditures by city range Kosovo — Mean per capita capital expenditures by city range 3.5 200 Capital Expenditures (Total – EUR) Capital Expenditures (Total – LEK) 3.0 2.5 150 2.0 100 1.5 1.0 50 0.5 0 0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 44 Key Trends and Findings Figure 33: Concentration of expenditures and population in the capital city 80% 70% 60% 50% 40% 30% 20% 10% 0% Albania BiH RS FBiH Croatia Kosovo Macedonia Montenegro Serbia Share of total expenditure Share of population Share of capital expenditure of these cities have additional public service deliv- However, by and large, capital cities’ budgets ery responsibilities given their importance in the reflect the population ratios. Beyond Macedonia, national spatial structure. Montenegro and Serbia, in the other countries, capital cities do not dominate the aggregate local Serbia and Macedonia are exceptional cases even budgets to the same extent. Their share in spend- among countries with the highest concentration. ing is proportional to their population size, which In Belgrade, local spending is more than two times is below one fifth of the country’s total population higher than the share of local residents: 44.6% of (except Podgorica). The census data are outdated in overall local government budget is spent in the some cases (Bosnia and Herzegovina, Macedonia), capital city, while only 20% of the population lives which might modify these proportions. in Belgrade. Skopje, in Macedonia, concentrates more than 70% of total capital expenditure in the More and often better services drive higher per country. Immigration from other countries of the capita expenditure in SEE capitals. Many specific former Yugoslavia might partially explain the con- urban services are available only, or to a much larger centration in the Belgrade. However, Serbia has a degree, in the capital cities, including public trans- balanced urban network with other major cities, portation and other main infrastructure services, e.g., Novi Sad, Nis, and Kragujevac, proportionally but also health and cultural services, including for located in the country, so the high concentration of citizens residing outside the capital, such as theaters, public funds is more likely a consequence of the hospitals, and museums. In Macedonia, most of the inherited urban services and the low level of fiscal local utility and communal service expenditures are equalization. Recently, the share of income tax real- concentrated in the capital (Skopje and its 10 com- located to local governments in Serbia was doubled prising municipalities). In Zagreb, the shares of local to 80%, which will further increase the differences spending on economic affairs (public transporta- among local governments. Personal income tax is tion), community amenities (water supply manage- primarily raised in the cities, so they will benefit the ment) and education are higher than the overall ratio most from these new PIT sharing rules, although the of the city budget in the country. Similarly, Tirana sharing ratio in Belgrade was set at a slightly lower dominates in the housing and community services level (70%). The large share of capital expenditure (typically water supply, street lighting), public in Skopje is likely to be explained by the large scale education and social protection. Prishtina is char- Government investment program in the capital city, acterized by a higher proportion of general public also known as “Skopje 2014”. services. Sarajevo is a special case, because most of 45 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 12: LOCAL PER CAPITA EXPENDITURES – 2011 Local Expenditures as % General government Local Expenditures of General Government Mean local expenditures expenditures in % of GDP as % of GDP Expenditures per capita 2011* EUR Albania 28.5% 2.2% 8.8% 48 BiH RS 49.2% NA NA 243 FBiH 191 Croatia 42.5% 6.8% 16.6% 459 Kosovo 30.0% 8.4% 48.6% 197 Macedonia 31.2.% 5.6% 20.4% 367 Montenegro 43.8% 3.9% 8.5% 506 Serbia 45.2% 6.8%* 13.7% 290* the municipal services are provided by the Sarajevo benefits and welfare programs were increased due Canton. Unfortunately, no reliable fiscal data were to the higher unemployment. available by functional classification from Belgrade. Public expenditure reductions affected sub- At the same time, capitals benefit from larger national capital and current budgets. Despite initial property tax revenues. Across the SEE region, capi- counter-cyclical measures to soften the impact of the tal cities collect the lion share of property related crisis, general government deficits and high public revenues, including property taxes and fees, and debt have forced most of the developed countries in communal charges levied on the property of busi- Europe to halt expanding fiscal policies. Spending on nesses or residents. Overall, the concentration of personnel, goods and services were cut back or they property related revenues in the capital city exceeds were reduced in real terms, by allowing the nominal both the share of population and the budget. They increases to be below the actual rate of inflation. benefit from the disproportionate concentration During the crisis, local governments were faced with of the tax base, but also typically higher property decreasing local property related tax revenues and values, which puts the large capital cities into a some of the local business taxes were also abolished favorable fiscal position. by the national economic recovery programs. As a reaction to the crisis, local governments had to 4.4  LOCAL GOVERNMENT REVENUE search for new options of revenue mobilization and efficiency gains in local spending.14 AND EXPENDITURE TRENDS DURING THE PERIOD 2008–2011 Local governments in SEE felt the downturn at different points in time. The 2008 economic and Sub-national governments could not avoid the financial crisis had a strong influence on public economic crisis’ fiscal implications. The economic finances since 2008 and continues determining fis- crisis had a strong influence on public finances since cal policies in the SEE region and beyond. From the 2008 and continues determining fiscal policies in data used for the MFR—which covers the period the SEE region and beyond. During 2008–2011, the 2008–2011—it is not possible to clearly differentiate trend in local finances was similar in most of the effects of the crisis from general structural trends. European countries. At the end of 2008 and in 2009, However, an analysis of the general revenue and the national governments tried to launch economic expenditure trends over this period suggests that stimulus programs. They targeted local capital investments and the local election cycle has also 14 Council of Europe, 2011: Local Government in critical times. Policies helped to keep capital expenditures at a relatively for crisis, recovery and a sustainable future. https://wcd.coe.int/ com.instranet.InstraServlet?command=com.instranet.CmdBlobGet high level. At the same time, spending on social &InstranetImage=2056216&SecMode=1&DocId=1873672&Usage=2 46 Key Trends and Findings Figure 34: Concentration of expenditures for different services in capital cities 90% 90% Share of expenditure by category in Capital city 80% 80% Share of population in capital city 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 30% 17% 20% 20% 14% 13% 20% 10% 7% 10% 0% 0% Albania BiH RS Croatia Kosovo Macedonia Montenegro General public services Education Environmental protection Economic a airs Housing and communal services Health Share of total population some local governments might have been hit ear- to municipalities might explain the delayed reaction lier by the fiscal restrictions and contracting eco- to the crisis. Capital cities followed the national nomic outputs. Already in 2009, local expenditures trends, although with greater fluctuation. dropped significantly in Montenegro (–18%) and in the RS (–8.6%); and decreased slightly in Serbia Large cities didn’t necessarily do better than their (–3.7%) compared to the previous year. In com- smaller peers. For example, Prishtina (Kosovo) parison, the crisis was delayed and less drastic in, and Tirana (Albania) were hit harder by the crisis e.g., Kosovo and Macedonia, but also Albania and than other local governments in their countries, Croatia. Particularly in Kosovo and Macedonia, local while Zagreb (Croatia), Podgorica (Montenegro) government expenditures continued to increase and Belgrade (Serbia) seem to have done bet- in real terms throughout this period, albeit with a ter during the first year of the crisis. Partly, this slower pace. In Albania, local expenditures were can be explained by the local revenue structure: still growing in 2009; and in Croatia, local expendi- Croatia and Serbia rely on shared revenues which tures did not decline—possibly due to the national automatically delayed the decline in resources. elections in 2010 and increased EU funding that In Montenegro, however, coastal municipalities, was made available to local authorities. However, which dominate total local expenditures, had a cutbacks were more serious later in 2010 (–9.8%). more sudden fall in property related revenues. In These trends are also reflected by the changes in Montenegro, where local expenditures fell drasti- local expenditures as share of in GDP. cally, at first, own revenues declined primarily in the Central region (outside Podgorica); and Those variations were driven mostly by different the shared PIT and property turnover tax in the local capital investment trends. Capital expenditure Southern coastal region. Later, after 2009, these was cut in Croatia, Montenegro and in the RS, while shared revenues partly recovered. The increase in it increased or remained stable in the other countries grants from the equalization fund compensated during the first year of the crisis. Albania followed the municipalities in the Central and the relatively this pattern with one year delay, while others could poor Northern regions. The election cycle has also keep (Kosovo) or even accelerate (Macedonia) capi- influenced capital city finances. For example, the tal spending by 2011. External sources of financing political stalemate in Tirana delayed council autho- for local capital investments that was made available rization of the budget in 2010. 47 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Municipalities in Montenegro and Macedonia Figure 36: Regional trends in local Governments sold assets or borrowed to cope with the revenue capital expenditures 2008–2011 gaps. In Montenegro, for example, local govern- ments have compensated revenue decrease with Regional Comparision – trends in capital expenditures asset sales in the coastal cities, or borrowing in the 380% poorer municipalities. Macedonian municipalities 330% were able to maintain their own source revenues 280% by following different policies in urban and rural 230% municipalities. The immediate reaction of rural 180% 130% municipalities was to increase local fees. They also 80% started to borrow from a very low base. Urban 30% municipalities and Skopje, on the other hand, had –20% the option to sell their assets. They also increased –70% 2008 2009 2010 2011 revenues from loans. Albania Croatia Kosovo Macedonia Montenegro Serbia In Croatia, the capital city experienced the most severe fall in own revenues. Among the recurrent revenues, administrative fees and charges declined in Zagreb starting 2010, while they only began falling government own revenues recovered relatively in 2011 in cities and municipalities. Capital revenues soon after their fall in 2010. Prishtina seem to have from sale of non-financial assets drastically declined been hit most by the economic downturn, causing in Zagreb in 2010, but quickly recovered in urban a sharp fall in own source revenues, while local local governments, although not in small municipali- governments in the rest of the country were able ties. Grants provided from the central following the to keep previous levels of own source revenues. general cuts in 2010 targeted primarily those smaller But local government budgets in Kosovo are municipalities, but interestingly, also the capital city. mostly financed by grants, which were stabilized at a lower level of growth, but still continued to Local revenues in Albania and Kosovo were sta- increase in 2010–2011. Albanian sub-national gov- bilized by transfers and grants. In Kosovo, local ernments faced restrictions with some delay since Figure 35: Regional trends in local Governments Figure 37: Local revenue trends in Montenegro expenditures 2008–2011 2008–2011 Regional Comparision — Evolution of Total Expenditures as share of GDP Montenegro (base 2007) 10% 130% 9% 8% 90% 7% 50% 6% 5% 10% 4% –30% 3% –70% 2% 1% –110% 0% 2008 2009 2010 2011 –150% 2008 2009 2010 2011 Albania FBiH** BiH RS** Croatia Own source Shared taxes Grants and Transfers Kosovo Macedonia Montenegro Serbia Borrowing Total Source: Data for BiH from aggregate data and correspondent to share of total BiH GDP 48 Key Trends and Findings unconditional transfers remained stable in 2009 for every year, reproducing a Gini coefficient analy- and later, the cuts were spread equally among the sis.15 For the second approach, each municipality is different types of local governments. Local own treated as an individual and a country is considered revenues—mainly small business tax and the tax as being highly unequal if the Gini coefficient is equal for building new infrastructure—started to decline to 1. The Gini coefficient is calculated for own source first at the intermediary tier (quark) and in Tirana revenues, total revenues, per capita own source (from 2010), partly caused by the new national revenues and per capita total revenues. regulations on capping user charges, fees and limiting municipal borrowing. Poorer municipalities could not consistently catch- up with their richer peers. Per capita expenditure Sub-national governments’ approaches to compen- trends reveal increased disparities between rich and sate for the revenue decline are likely to affect their poor municipalities in some countries, and conver- future position. Different responses across the SEE gence in others. In Kosovo, poorest municipalities region provide some general lessons as how local have achieved to close the gap to richer municipali- governments were able to cope with the economic ties in their per capita expenditures (see Figure 38). crisis. Cities with marketable assets chose to increase In Macedonia, on the other hand, poorer municipali- asset sales as the immediate reaction to the crisis. ties have not achieved to close the gap. Here, per For example, in Macedonia, but also in Croatia capita expenditure increased overall in parallel in first only in Zagreb, but recently in all other cities. the richest and poorest municipalities, despite with Middle tier local governments in Albania lost their a slightly diverging trend. Worryingly, the spending own source revenues, so probably they lacked the difference between poor and rich municipalities real incentives to go after these diverse and minor grew considerably in Montenegro (see Figure 39). resources. Borrowing was a new option, which was In Croatia, disparities were maintained during used extensively first in rural, later in the urban 2008 and 2011, and Albania experienced a slight municipalities of Macedonia. increase in the gap between the poorest and the richest after 2010. 4.5  INEQUALITY, EQUALIZATION, Spending patterns differ remarkably in munici- AND SELECTED TRENDS BY TYPES palities with high and low per capita expenditure. OF MUNICIPALITIES In Kosovo, while there is convergence in per capita expenditures, poor and rich municipalities have 4.5.1  Trends of increasing disparities and very different spending priorities. Poor municipali- convergence ties increased considerably their expenditures on health and general public services while, during This subsection contains an analysis of inequalities the same period, rich municipalities increased their looking at disaggregated data. The main purpose expenditures on environmental protection and of this analysis is to review existing revenue and housing. During the same period, expenditures expenditures inequalities among municipalities and from other categories, such as defense and public assess trends during the period 2008–2011. We use order, were reduced (see Figure  40). In Albania, two approaches to evaluate trends in inequalities. The first approach consists of a quintile analysis 15 The Gini coefficient is a measure of statistical dispersion and mea- in which municipalities are regrouped according sures the inequality among values of a distribution. For this exer- cise we calculate four different Gini coefficients using the following to their per capita own source revenue in quintiles variables: (1) per capita own source revenue, (2) per capita total revenue, (3) own source revenue and (4) total revenue. The first two using 2008 as a reference year. Municipalities are estimations are useful to evaluate if municipalities across the coun- considered “rich” within a country if they belong try have access to the same level of resources—taking into account their population—and if grants and transfers are playing a role in to the fifth quintile; and “poor” if they belong to a reduction on the level of inequality. The second two estimations the first quintile. The second approach consists of provide evidence on how municipal revenues are distributed in a country and might serve as a proxy of the level of urban primacy of an analysis of the per capita revenue distribution a given country. 49 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans In Croatia and Macedonia, results suggest a slight Figure 38: Converging trend in per-capita reduction in per capita own-source revenue inequali- expenditures for poorest (1st Quintile) and richest ties during 2008 and 2011. Total per capita revenue (5th Quintile) – 2008 reference year inequalities grew in RS, Kosovo and Montenegro, but Kosovo were reduced in Albania, Croatia and Macedonia. 140 However, for all years and countries, except for the Mean per capita expenditure per 120 RS in 2009, the level of per capita revenue inequal- 100 ity is greater when computing own source revenues revenue quintile 80 as opposed to general revenues, suggesting that 60 equalization grants are working. The regional trends on per capita revenue inequalities using the Gini 40 coefficient methodology can be found in Figure 42. 20 0 Own-source revenue inequality increases slightly 2008 2009 2010 2011 1 5 in Albania and Kosovo. All of the other coun- tries experienced a reduction in the level of own source inequality during the period 2008–2011 (see Figure 43). An interesting result is that the level of richest municipalities increased considerably their own source revenue inequalities is much higher expenses in housing and “others” category while than the level of per capita own source revenue achieving to maintain their expenditures for almost inequalities. This finding suggests that, while there all other categories. Poorest municipalities, on the is a large concentration of resources in a number contrary, experienced a reduction in most of the of municipalities, these resources are more or less expenditures categories with social protection and proportionate to the population living there. Total health being the most affected sectors. revenue inequalities also grew in Albania, but were reduced or maintained in the other countries. In Per capita own-source revenue inequality increases addition, for all years and countries, the level of in Albania, Kosovo, Montenegro, and the RS. revenue inequality is greater when computing own Findings suggest a visible growth of per capita own- source revenues as opposed to general revenues, source revenue inequality during the period of analy- which in turn suggests again that equalization sis, measured by the Gini coefficient methodology. grants seem to be working as intended. Figure 39: Diverging trends in per-capita expenditures for poorest (1st Quintile) and richest (5th Quintile) – 2008 reference year Macedonia Montenegro 35,000 1,600 Mean per capita expenditure per Mean per capita expenditure per 30,000 1,400 25,000 1,200 revenue quintile revenue quintile 1,000 20,000 800 15,000 600 10,000 400 5,000 200 0 0 2008 2009 2010 2011 2007 2008 2009 2010 2011 1 5 50 Key Trends and Findings Figure 40: Trends in functional expenditures for poorest (1st Quintile) and richest (5th Quintile) in Kosovo – 2008 reference year Kosovo — QUINTILE 1 (base 2008) Kosovo — QUINTILE 5 (base 2008) 1000% 1000% 900% 900% 800% 800% 700% 700% 600% 600% 500% 500% 400% 400% 300% 300% 200% 200% 100% 100% 0% 0% –100% –100% 2008 2009 2010 2011 2008 2009 2010 2011 General public services Defense Public order Environment Housing Health Economic a airs Education Social protection Recreation culture and religion 4.5.2  Equalization of expenditure needs directed towards the equalization of expenditure and fiscal capacity needs or fiscal capacity. Differences in per capita grants may reflect efforts to equalize the underly- Overall, the general grant transfers, as utilized to ing differences in expenditures needs arising from an increasing extent, seem to meet equalization differences in size and composition of population objectives across the SEE region. Unconditional and other, similar expenditure drivers. As appears budget transfers from central government may be from Table 13, generally, local governments with Figure 41: Trends in functional expenditures for poorest (1st Quintile) and richest (5th Quintile) in Albania – 2008 reference year Albania — QUINTILE 1 (base 2008) Albania — QUINTILE 5 (base 2008) 530% 530% 460% 460% 390% 390% 320% 320% 250% 250% 180% 180% 110% 110% 40% 40% –30% –30% –100% –100% 2008 2009 2010 2011 2008 2009 2010 2011 General public services Education Environment Housing Health Economic a airs Social protection Recreation culture and religion Other 51 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 42: Regional trends of Local Governments’ Gini coefficient – per capita Total Revenue 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Albania FBiH BiH RS Croatia Kosovo Macedonia Montenegro 2008 2009 2010 2011 the highest amount of per capita transfer received an important, although not equal to grants, fund 1.5–3.8 more than the average in the respective sources in the smaller municipalities. Expenditure group of cities or municipalities in the majority variations and levels (expenditure needs), by size of SEE countries. In addition, the level of grant of municipalities vary across the countries—in per capita is decreasing by size of municipalities, Kosovo the per capita expenditure is decreasing reflecting the efforts to compensate for low fiscal by size of municipalities, implying that per capita capacity/higher expenditure needs in the smaller, spending in the largest municipalities amount to often rural municipalities, which may not be in a less than 50% of per capita spending in the smallest position to afford service delivery without raising municipalities. Somewhat similar trend is observed tax efforts excessively. in Montenegro. In Albania and Macedonia, on the other hand, while the per capita expenditure does Per Capita Expenditure, own sources, and grant not differ significant across the various sizes of transfers by size of municipalities vary signifi- municipalities in the two countries, a minor increase cantly across countries. Some highlights (see Table by size is observed in Albania, with expenditure 14): In Albania and Kosovo, the per capita grant spend of +20% in the largest municipalities. Finally, is by all measures the dominant fund source in in Croatia, a somewhat similar pattern to Albania the smaller municipalities, while in Montenegro is observed, although the expenditure is much and Macedonia, own fiscal efforts per capita is higher in the larger municipalities than in smaller. Figure 43: Regional trends of Local Governments’ Gini coefficient – per capita Own Source Revenue 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Albania FBiH BiH RS Croatia Kosovo Macedonia Montenegro 2008 2009 2010 2011 52 Key Trends and Findings Figure 44: Regional trends of Local Governments’ Gini coefficient – Own Source Revenue 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Albania FBiH BiH RS Croatia Kosovo Macedonia Montenegro 2008 2009 2010 2011 These patterns in expenditure needs are reflected municipalities differently in the SEE countries. in the required need for ‘gap filling’ by the gen- The following section provides selected insights in eral grants from central government. In Kosovo, different revenue and expenditure trends by type of where expenditure needs per capita were highest municipalities. The disaggregated data set collected in the smaller municipalities, the grant transfers, in by the MFR permits a more granulated analysis and absolute and as percentages, provide a core part of allows for modeling the effects of policy decisions the funding, probably due to relatively low fiscal on different sub-sets of municipalities. The selected capacity. In Macedonia and Croatia, on the other examples discussed in the following should be hand, we do not see the same important role of the interpreted in a context of demonstrating analytical grant funding, in general, but also in the smaller tools of the MFR data set, rather than as a basis for municipalities. specific policy recommendations. To make specific suggestions for policy would require a more in- 4.5.3  Selected trends: who benefitted, who depth review of the variables influencing the trends fell behind? discussed below. The local government finance architecture and Total revenue trends suggest large variations across intergovernmental fiscal relations has affected municipalities of different size in Albania and Figure 45: Regional trends of Local Governments’ Gini coefficient – Total Revenue 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Albania FBiH BiH RS Croatia Kosovo Macedonia Montenegro 2008 2009 2010 2011 53 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans TABLE 13: PER CAPITA GENERAL GRANTS BY SIZE OF LOCAL GOVERNMENT UNIT 2011 (FBIH: 2010) Size of Local Government Unit – EURO less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 more than 200000 Albania 109.2 63.2 56.3 56.0 64.8 36.4 BiH RS 119.8 21.3 15.5 6.1 4.6 FBiH 413.0 89.7 77.8 58.0 54.5 Croatia 106.9 78.0 65.8 68.6 36.4 10.5 Kosovo 364.9 230.8 175.2 166.9 156.3 124.4 Macedonia 107.4 113.5 99.4 98.9 128.0 41.9 Montenegro 359.4 220.5 87.3 56.8 45.0 TABLE 14: PER CAPITA EXPENDITURE, OWN SOURCE, GRANTS AND TRANSFERS AND TOTAL REVENUE BY SIZE OF LOCAL GOVERNMENT UNIT 2011 ALBANIA – Size of Local Government Unit – EURO 100000– more than PER CAPITA less than 5000 5000–10000 10000–40000 40000–100000 200000 200000 Average Expenditure 50.6 41.4 48.9 59.0 62.5 67.5 55.0 Own source revenue 3.4 9.7 21.4 33.0 26.4 36.7 21.8 Grants and transfers 109.2 63.2 56.3 56.0 64.8 36.4 64.3 Total revenue 112.4 72.8 77.8 89.4 91.9 73.0 86.2 KOSOVO – Size of Local Government Unit – EURO 100000– more than PER CAPITA less than 5000 5000–10000 10000–40000 40000–100000 200000 200000 Average Expenditure 372.3 235.6 196.0 185.2 182.6 165.4 222.8 Own source revenue 21.4 9.4 29.0 20.6 26.8 52.0 26.5 Grants and transfers 364.9 230.8 175.2 166.9 156.3 124.4 203.1 Total revenue 386.3 240.1 198.4 187.4 183.1 176.4 228.6 CROATIA – Size of Local Government Unit – EURO 100000– more than PER CAPITA less than 5000 5000–10000 10000–40000 40000–100000 200000 200000 Average Expenditure 466.8 396.5 464.2 553.2 663.0 1148.4 508.8 Own source revenue 353.8 317.8 387.3 486.0 566.2 1107.7 536.5 Grants and transfers 106.9 78.0 65.8 68.6 36.4 10.5 61.0 Total revenue 460.1 394.9 453.1 554.6 602.5 1118.2 597.2 MACEDONIA – Size of Local Government Unit – EURO 100000– more than PER CAPITA less than 5000 5000–10000 10000–40000 40000–100000 200000 200000 Average Expenditure 399.4 394.0 316.4 419.2 398.5 295.8 370.5 Own source revenue 59.2 57.5 39.2 76.9 58.6 82.3 62.3 Grants and transfers 107.4 113.5 99.4 98.9 128.0 41.9 98.2 Total revenue 164.8 174.3 142.7 182.3 188.0 132.8 164.1 MONTENEGRO – Size of Local Government Unit – EURO 100000– more than PER CAPITA less than 5000 5000–10000 10000–40000 40000–100000 200000 200000 Average Expenditure 1417.8 348.8 386.3 263.1 289.8 — 541.1 Own source revenue 896.3 87.8 244.0 144.4 193.1 — 313.1 Grants and transfers 359.4 220.5 87.3 56.8 45.0 — 153.8 Total revenue 1693.6 412.4 421.3 274.2 291.4 — 618.6 54 Key Trends and Findings Figure 46: Changes in total revenues by size of municipalities (Albania and Montenegro) Albania — changes General Revenue by city size Montenegro — changes General Revenue by city size 100% 240% 75% 170% 50% 100% 25% 30% 0% –40% –25% –50% –110% –75% –180% –100% –250% 2008 2009 2010 2011 2008 2009 2010 2011 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Montenegro. While, overall, down- and upward Figure 47: Changes in total revenues, own source, trends in total local revenues show comparably and grants and transfers (urban vs. rural) in similar patterns in the municipalities of BiH, Croatia, Macedonia Kosovo, and Macedonia; trends vary significantly for municipalities of different size in Albania and Macedonia — changes in General revenue urban vs rural Montenegro. In Albania, we find overall stagna- 58% tion of total revenues across municipalities after an 38% increase from 2009–2010. However, municipalities 18% with a population between 100,000–200,000, MFR –2% data indicate a strong decline in total local rev- –22% –42% enues. In Montenegro, on the other hand, smaller –62% municipalities have increased their total revenues; 2008 2009 2010 2011 in particular municipalities with less than 5,000 Macedonia — changes in own source revenue urban vs rural inhabitants registered a steep increase in total rev- enues after 2009. 58% 38% Revenue trends in Macedonia suggest rural 18% municipalities outperform their urban peers in –2% mobilizing revenues. During 2008–2011, rural –22% municipalities’ total revenues grew faster than –42% urban municipalities total revenues. In addition, –62% 2008 2009 2010 2011 rural ones seem to have managed offsetting an initial decrease in own source revenues. However, Macedonia — changes in Grants and Transfers urban vs rural a closer review of the data reveals that rural municipalities’ better performance in total revenue 80% mobilization was largely the result of increasing 60% grants and transfers from the central government 40% (see Figure 47). 20% Skopje has the fastest increase in expenditure 0% 2008 2009 2010 2011 despite slower revenue growth. Another observa- Urban Rural tion from the disaggregated data suggests very 55 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans different reactions on the expenditure side given long-term or multi-year investment programs, the diverse revenue trends in municipalities of including those financed from loans and grants, or different size and economic weight. In Macedonia, deliberate counter-cyclical investments to provide for example, Skopje as the capital city and larg- a fiscal stimulus to the local economy. However, est municipality experienced the slowest revenue the diverging revenue and expenditure trends also growth during the observed time period. However, require close monitoring to avoid fiscal stress as total expenditure rose fastest in that city (see a result of unbalanced budgets over an extended Figure 48). A possible explanation are the ongoing period of time. Figure 48: Revenue and expenditure trends by size of municipalities (Macedonia) Macedonia — changes General Revenue by city size Macedonia — changes total expenditures by city size 800% 300% 600% 250% 400% 200% 200% 0% 150% –200% 100% –400% –600% 50% –800% 0% 2008 2009 2010 2011 2008 2009 2010 2011 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 56 1 5 Conclusion and Directons for a Possible Way Forward This report has demonstrated the increasing relative weight and role of sub-national governments in the SEE region. In line with their increas- ing importance in public service delivery and spending over the past decade, the way municipalities raise, allocate, and manage their finan- cial resources becomes increasingly important. Greater fiscal authority and responsibilities drive a need for improved planning, budgeting and financial management—both at the local and central level. Reliable and high quality municipal finance data is critical in this context. However, the public finance capacity of municipalities in the SEE Region remains insufficient and only partially equipped with the appropriate tools, techniques and data. Public finance processes and procedures are viewed as institutional requirements imposed by central authorities, and further compounded by the lack of data and financial management techniques, the local authorities lack the incentives to set directions, pri- oritize and ensure efficiency in public service delivery and expenditure. In this context, this report has aimed to address identified capacity issues by the creation of a regional dataset for disaggregated municipal finance analysis, with the goal to contribute to improved understand- ing of local government management and finance in the SEE Region; and improving the quality and consistency of key municipal finance data for improved evidence based policy making. Although limited in scope and expandable in depth, MFR has piloted disaggregated data collection and analysis for in-country analyses and regional compari- son in SEE. At the outset of the analytical work and capacity building activity, knowledge and analysis of municipal finance issues in SEE was limited mostly to aggregated, national datasets. With the support from regional and local experts and the Local Government Associations in the participating entities, MFR has established a regional municipal finance database available for further analysis. Overall, the database meets expectations as regards coverage of countries, variables and time period, and, as evidenced in the report, has provided a very strong foundation for municipal finance analyses. 57 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans A number of data gaps and data quality issues per- to sustain regional and country-specific datasets sist. These include issues typical for municipal data- and analytical capacity within SEE countries for sets and which are outside the control of this project, continued analysis, including outreach to national including but not limited to blurred definitions of authorities on data and analysis? How to establish assigned functions, poor quality of recording prac- further capacity of local associations in each country tices on expenditures and revenue classifications, to maintain, expand and update the database? How unclear definitions on administrative boundaries, would a regional database be maintained, and by and practical difficulties in applying concepts of whom? How to broaden the concept of a database ‘own source revenues’, among others. Other data and analytical review and include other countries gap and issues, however, may be attributed to the in the ECA region? project as such, and efforts should be undertaken to address these going forward. This includes effort to A number of issues would need to be addressed to improve data coverage, including coverage of identi- respond to these questions with the goal to (i) make cal periods of time, across countries. municipal finance data collection regular; and (ii) develop a user friendly system of information man- However, the potential use and application of disag- agement. The following summarizes suggestions gregated data go beyond the initial analysis possible and recommendations for a possible way forward: within the scope and limitations of the MFR. For example, compared to aggregated national datasets, Regular transfer of detailed municipal finance data disaggregated sub-national finance data allows for and information. In all countries in the SEE region, benchmarking across municipalities. Municipal cur- disaggregated financial reports on local govern- rent and capital expenditures analyzed together with ment expenditures and revenues are available at service performance indicators and statistics may pro- the Ministry of Finance or at the Treasury. The only vide opportunities for comparing facility unit costs, exception is the Federation of BiH, where the local but also to reviewing variations in service delivery government association with the support from the outcomes against different expenditure needs; spend- MFR has launched a survey to collect municipal fis- ing- and revenue trends. However, service efficiency cal data (see Box 2). The survey format was similar measures are difficult to compare across countries. to the financial reports used by local governments. Different functional assignments for the same type In the first year, the response rate was rather high of sub-national entities make comparisons less pre- around 95% of total local budgets. This survey, cise; fiscal information on the evaluated service is supported by the MFR expert, is the only source of not always sufficiently detailed; the precise scope of individual LGU detailed fiscal data for the Ministry service provided differs in practice; and municipal of Finance, and it was agreed with the LGA to con- accounting practices follow different rules. Although tinue regular data collection on that basis. these limitations reduce the scope for regional bench- marking, a concerted effort on service outcome data In the BiH RS, with the support of the MFR, the local collection comparable to the MFR can overcome main government association was able to collect disag- constraints and further deepen our understanding of gregated data from the Ministry of Finance for the local service delivery performance in SEE. first time. In Kosovo and Macedonia, detailed finan- cial data are produced by the financial accounting By and large, constructing a consolidated regional and reporting systems. The ‘FreeBalance’ software database on municipal finance in the SEE coun- provides sufficient information in Kosovo, which tries and the related analysis provides significant can be re-grouped according to the LGA needs. In value-added to the knowledge and understand- Macedonia the very detailed data should be also ing of the policy challenges facing municipalities recoded and transformed for analytical purposes. in these countries. However, going forward, a number of questions for follow-up actions may In Albania, Croatia, Montenegro and Serbia, the national be considered. These include, among others: How reporting system was able to provide municipal 58 Conclusion and Directons for a Possible Way Forward finance data for use in the MFR. Only minor trans- formation was needed on these disaggregated fis- Box 2: The implementation and application of the MFR cal data. Unfortunately, data was not requested by in the Federation of BiH the LGA in Serbia and could not be included in the In the Federation of BiH, implementation of the MFR encompassed detailed, disaggregated analysis, but is available in a wider scope of activities, due to the fact that the Federal Ministry principle at the MoF/Treasury. of Finance (FMF) does not collect the data from individual LGUs (FMF collects LGU data grouped by Cantons). Neither had the Asso- A formal agreement between the local govern- ciation of the Municipalities and Cities of FBiH have comprehensive ment association and the responsible govern- fiscal data needed for the MFR exercise. ment agency would help facilitating regular data Consequently, data had to be collected through a survey with transfer. In Montenegro, for example, the Union of requested table templates sent to each of the 80 local government Municipalities (UoM) has agreed to propose to the units (LGUs) in the FBiH through the Association of the Municipali- ties and Cities of FBiH. In order to minimize the additional work Ministry of Finance signing of a Memorandum of required by the LGUs, existing reporting templates which they Understanding for regular data transfer from the have to fill out for other purposes were used (reporting templates Ministry of Finance Budget Department to the UoM which the municipalities sent to Cantons, FMF or tax authorities). on a quarterly basis. Total of five tables (with comprehensive disaggregated data on: revenues, expenditures and financing; balance sheet; functional Construction of consolidated municipal finance expenditure categories; debt stock; and current debt categories) were prepared and sent to each LGU in FBiH through the Associa- data sets: data clean-up and re-categorization. tion in Spring 2012. The Association (with World Bank technical Transformation of the datasets in line with the LGA assistance) intensively worked on contacting LGUs to deliver data needs is the next step after receiving data from the and within the data cleaning process, which spanned until late national government. As data has already been Fall 2012. In total, data was received from 72 LGUs, out of total 80 verified by the relevant ministry, there is no need to LGUs, covering around 95% of total LGU expenditures in FBiH. make any corrections of possible errors at this stage. The Association has already used the MFR database for various However, in the case of FBiH, the survey might purposes, including: as inputs for NALAS reports in fiscal decentral- produce some false data, for example by different ization; for internal and external discussions about the framework of revenue distribution and budgetary legislation in FBiH; and for reporting, e.g., in thousands, instead of millions of discussions about LGU finances and reporting with international local currency. Other sources of error might include, organizations such as the IMF. for example, that local governments fail to follow In addition, with the goal of achieving sustainability of the data uniform accounting practices. Generally, these collection process by the Association and regular update of the MFR issues do not occur if data is transferred from the database, the Association has issued another request for 2012/2013 national reporting system, which usually addresses data to all LGUs in Spring of 2013 (within this round of data collec- these issues already. tion it is also expected that the Association will collect the previous data for the 8 municipalities which did not send data last year). Data presentation. Disaggregated municipal Given that the Association is currently the only institution in FBiH which has desegregated detailed data for individual municipalities, finance data can be used for various purposes. FMF has asked the Association to share the MFR database. Among others, the goal of municipal finance data and analysis is to inform local government officials, Source: Naida Carsimamovic. investors, media, civil society organizations and the general public about the financial position and overall performance of a municipality. This requires public access to selected information, or to indica- developed by either national governments or statis- tors, depending on the level of available municipal tical offices, e.g., in Denmark (http://www.noegle- finance data. Publicly accessible databases can also tal.dk/). However, non-governmental organizations offer search options, such as aggregation by regions, and research groups also provide comparable tools. comparison of local governments, or the averages For example, in Sweden, the local government for the a subset of municipalities (e.g. similar size, association and national government jointly oper- administrative status). Such services are usually ate a municipal finance database (http://www. 59 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans kolada.se/). A non-profit organization took the the risk of limited sustainability. ZELS, in Macedonia, responsibility for a similar database in Portugal owns a sophisticated municipal finance database and (http://www.pordata.pt/en/Municipalities/ analytical tool, developed with assistance from the Search+Environment/Table). A similar service is USAID supported Macedonia Local Government available in Macedonia, developed by a think tank Activity (MLGA) project. However, regular updat- in collaboration with the MFR local expert (http:// ing and maintenance lacks sustainable funding since www.mkbudget.org/els/). the project closed in 2011, and alternative financing solutions need yet to be developed. As a result, Municipal finance reports can also be combined methods of regular data-collection and management with other data, which would allow the calculation have to be developed by or in close cooperation with of indicators by users. They can be reported in a the partner host organization. Related costs of data more sophisticated way, using visual presentations, collection and maintenance need to be planned and for example connecting financial indicators to geo- budgeted upfront to ensure sustainability. Since local graphic information creating maps. In Croatia, the government associations in the target countries are Association of Cities already started to use MFR in different positions with regard to their mandates, data and decided to develop a geo-visual form of organizational, and staff capacities, business models presentation. Figure 49 provides an example of local for long term sustainability necessarily will have to government labor costs per capita combined with be different. Even if interested LGAs want to raise uncollected current revenues per capita. revenues by operating a municipal finance database, it would be advisable to develop an open component Data management capacity. Generally, local govern- of the indicators, which is accessible free of charge. ment associations are not prepared for the complex This public information will generate demand for task of managing municipal finance data and present the fee based services. the generated information in an accessible and user- friendly format. LGAs may employ fiscal experts and communication managers, but that would require Figure 49: Visual presentation of municipal close cooperation between these types of expertise. finance indicators in Croatia In the past, LGAs have not had a business model for developing and operating municipal finance data and typically relied on donor support. For example, the LGAs in Albania, Montenegro, and the Federation of BiH have expressed their interest in leveraging donor financial assistance for regular municipal finance data management. At the regional level, if demonstrated demand exists, NALAS would be a natural host for a regional municipal finance database. Regional analysis of country aggregate data can be based on national datasets, while for more detailed analysis, NALAS may use information produced by disaggregate databases in their member associations. Funding for regular maintenance and sustainabil- Source: Croatia Association of Cities. ity. However, external financing and assistance bears 60 1 6 References Bajo, A. (2012), Municipal Finance Review of Croatia. MFR Input Report. Brdarevic, L. (2012), Municipal Finance Review of Serbia. MFR Input Report. Carsimamovic, N. (2013), Municipal Finance Review of the Entity of Federation of Bosnia and Herzegovina. MFR Input Report. Council of Europe, 2011: Local Government in critical times. Policies for crisis, recovery and a sustainable future. [https://wcd.coe.int/com. instranet.InstraServlet?command=com.instranet. CmdBlobGet&In stranetImage=2056216&SecMode=1&DocId=1873672&Usage=2] Frank, Jonas and Martinez-Vazquez, Jorge (eds.). 2013. “Decentralization and Infrastructure: From Gaps to Solutions”. Georgia State University and World Bank, forthcoming. Gabor, P. (2013), Local government finances in SEE: Regional Review of the Western Balkans. MFR Input Report. Gjika, A. (2012), Municipal Finance Review of Albania. MFR Input Report. IMF (2013), IMF World Economic Outlook Data: Hope, Realities, Risks. Ministry of Finance of the Republic of Serbia (2012), Public Finance Bulletin 2008–2011. NALAS (2012), Report of Fiscal Decentralization Indicators for South- East Europe 2006–2011. Nikolov, M. (2012), Municipal Finance Review of FYR of Macedonia. MFR Input Report. Obradovic, N. (2012), Municipal Finance Review of Montenegro. MFR Input Report. Osman, S. (2012), Municipal Finance Review of Kosovo. MFR Input Report. Rakic, G. (2012), Municipal Finance Review of the Entity of Republika Srpska of Bosnia and Herzegovina. MFR Input Report. World Bank (2012), ECA FACE database. World Bank (2013), Serbia Municipal Finance and Expenditure Review 61 Annexes Annex 1: Summary Overview Graphs 65 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 50: Variations in total revenue sources by city size (As percentage of total) Albania Croatia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Kosovo Macedonia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Own source Shared taxes Grants and Transfers Borrowing Other (continued on next page) 66 Annex 1: Summary Overview Graphs Figure 50: Variations in total revenue sources by city size (As percentage of total) (continued) Montenegro BiH RS 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 40000–100000 100000–200000 10000–40000 FBiH 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Own source Shared taxes Grants and Transfers Borrowing Other 67 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 51: Variations in total revenue sources by city size (Absolute values) Albania Croatia 16,000,000 7.00E+09 14,000,000 6.00E+09 12,000,000 5.00E+09 10,000,000 4.00E+09 8,000,000 3.00E+09 6,000,000 4,000,000 2.00E+09 2,000,000 1.00E+09 0 0.00E+00 More than 200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 Kosovo Macedonia 1.60E+08 1.20E+10 1.40E+08 1.00E+10 1.20E+08 1.00E+08 8.00E+09 8.00E+07 6.00E+09 6.00E+07 4.00E+09 4.00E+07 2.00E+07 2.00E+09 0.00E+00 0.00E+00 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Montenegro BiH RS 1.20E+10 2.50E+08 1.00E+10 2.00E+08 8.00E+09 1.50E+08 6.00E+09 1.00E+08 4.00E+09 2.00E+09 5.00E+07 0.00E+00 0.00E+00 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Own source Shared taxes Grants and Transfers Borrowing Other (continued on next page) 68 Annex 1: Summary Overview Graphs Figure 51: Variations in total revenue sources by city size (Absolute values) (continued) Bosnia Fed 300,000 250,000 200,000 150,000 100,000 50,000 0 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 more than 200000 Own source Shared taxes Grants and Transfers Borrowing Other 69 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 52: Variations in own source revenue sources by city size (as percentage of own source) Albania Croatia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 100000–200000 More than 200000 100000–200000 More than 200000 40000–100000 40000–100000 Less than 5000 5000–10000 10000–40000 Less than 5000 5000–10000 10000–40000 Kosovo Macedonia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Montenegro BiH RS 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Taxes Fees and charges Rents Fines Capital Other 70 Annex 1: Summary Overview Graphs Figure 53: Variations in own source revenue sources by city size (Absolute values) Albania Croatia 4.50E+06 2.50E+09 4.00E+06 3.50E+06 2.00E+09 3.00E+06 1.50E+09 2.50E+06 2.00E+06 1.00E+09 1.50E+06 1.00E+06 5.00E+08 5.00E+05 0.00E+00 0.00E+00 40000–100000 100000–200000 More than 200000 40000–100000 100000–200000 More than 200000 10000–40000 10000–40000 Less than 5000 5000–10000 Less than 5000 5000–10000 Kosovo Macedonia 40,000,000 5.00E+09 35,000,000 4.50E+09 4.00E+09 30,000,000 3.50E+09 25,000,000 3.00E+09 20,000,000 2.50E+09 15,000,000 2.00E+09 1.50E+09 10,000,000 1.00E+09 5,000,000 5.00E+08 0 0.00E+00 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Montenegro BiH RS 6.00E+07 1.80E+08 1.60E+08 5.00E+07 1.40E+08 4.00E+07 1.20E+08 1.00E+08 3.00E+07 8.00E+07 2.00E+07 6.00E+07 4.00E+07 1.00E+07 2.00E+07 0.00E+00 0.00E+00 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Taxes Rents Capital Fees and charges Fines Other 71 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 54: Functional expenditures by city size (As percentage of total expenditures) Albania Croatia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 100000–200000 40000–100000 More than 200000 100000–200000 40000–100000 More than 200000 Less than 5000 5000–10000 10000–40000 Less than 5000 5000–10000 10000–40000 Kosovo Macedonia 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Montenegro 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 General public services Defense Public order Social protection Economic a airs Environment Housing Recreation culture and religion Health Education Other 72 Annex 1: Summary Overview Graphs Figure 55: Functional expenditures by city size (Absolute values) Albania Croatia 10,000,000 1.20E+10 9,000,000 8,000,000 1.00E+10 7,000,000 8.00E+09 6,000,000 5,000,000 6.00E+09 4,000,000 3,000,000 4.00E+09 2,000,000 2.00E+09 1,000,000 0 100000–200000 0.00E+00 More than 200000 100000–200000 40000–100000 40000–100000 More than 200000 Less than 5000 5000–10000 10000–40000 Less than 5000 5000–10000 10000–40000 Kosovo Macedonia 1.60E+08 1.40E+10 1.40E+08 1.20E+10 1.20E+08 1.00E+10 1.00E+08 8.00E+09 8.00E+07 6.00E+09 6.00E+07 4.00E+07 4.00E+09 2.00E+07 2.00E+09 0.00E+00 0.00E+00 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 More than 200000 Montenegro 6.00E+07 5.00E+07 4.00E+07 3.00E+07 2.00E+07 1.00E+07 0.00E+00 Less than 5000 5000–10000 10000–40000 40000–100000 100000–200000 General public services Defense Public order Social protection Economic a airs Environment Housing Recreation culture and religion Health Education Other 73 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 56: Per capita expenditure in Environmental Protection, Education and Housing against population density – Croatia, Kosovo and Macedonia Density analysis (Croatia – Kosovo – Macedonia) 1 line Environment capita, 2 line housing, 3 line education Kosovo Macedonia 1.5 1,500 1.0 1,000 exp_env_cap exp_env_cap 0.5 500 0 0 0 1000 2000 3000 4000 0 5000 10000 15000 lgu_density lgu_density 60 10,000 8,000 40 exp_hous_cap exp_hous_cap 6,000 4,000 20 2,000 0 0 0 1000 2000 3000 4000 0 5000 10000 15000 lgu_density lgu_density 80 10,000 8,000 60 exp_edu_cap exp_edu_cap 6,000 40 4,000 20 2,000 0 0 0 1000 2000 3000 4000 0 5000 10000 15000 lgu_density lgu_density (continued on next page) 74 Annex 1: Summary Overview Graphs Figure 56: Per capita expenditure in Environmental Protection, Education and Housing against population density – Croatia, Kosovo and Macedonia (continued) Density analysis (Croatia – Kosovo – Macedonia) 1 line Environment capita, 2 line housing, 3 line education Croatia 4000 3000 exp_env_cap 2000 1000 0 0 1000 2000 3000 lgu_density 1000 800 exp_health_cap 600 400 200 0 0 1000 2000 3000 lgu_density 4000 3000 exp_edu_cap 2000 1000 0 0 1000 2000 3000 lgu_density 75 South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans Figure 57: Diverging and Converging trends of per capita expenditures between “rich” and “poor municipalities” Inequalities graphs for each of the countries Albania Croatia 10,000 6 8,000 5 4 6,000 3 4,000 2 2,000 1 0 0 2008 2009 2010 2011 2008 2009 2010 2011 Mean per capita expenditure per revenue quintile Kosovo Montenegro 140 1,600 120 1,400 1,200 100 1,000 80 800 60 600 40 400 20 200 0 0 2008 2009 2010 2011 2007 2008 2009 2010 2011 Macedonia 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 2008 2009 2010 2011 1 5 76 Annex 2: Data Gap Analysis and Reclassification 77 B&H Country Albania Albania FBiH RS Croatia Kosovo Macedonia Montenegro Serbia Number of local 80 but only 53 with 573 (2011), 576 government units reviewed data only (2010), 576 (2009), 170* (More with providing data 385 386 for 2010 63 576 (2008) 38 85 21 different regions) Years covered for analysis 2008–2011 2008–2012 2010 2008–2010* 2008–2011 2008–2011 2008–2011 2007*-2011 2008–2011** Major gaps in the Data on transfers Not all municipalities Inconsistent Data on borrowing Expenditure For expenditures Information on Database is very dataset from Min Interior submitted data and categories among has not been classification is by programs there expenditures limited on both with some issues some data had errors years. The first merged with divided for each is an inconsistency by economic revenues and for year 2011. and was sent back (no stage database panel. Data category among in the names of the category is very expenditures. Not clear from correction of errors was only contains on statistics Expenditure variables for 2008 complete from Expenditures database if “-” received). Uses OCDE 2008–2010. dissagragated by economic (M and V variables) 2007–2011, only available for means missing expenditure functional Expenditures very complete categories. Capital agains the rest of by functional a selected set of or zero. Missing classification. The by functional with additional expenditure the years. National categories is very categories and variables in 2010 database has serious categories are indicators (i.e. classification Statistics contain incomplete (a lot until 2010. (Tax vehicles), and datagaps for all missing. Data on development is very good information on of cities missing 2011 (tax vehicles, variables. population and index, area, etc). and contains inflation, capital info) and only from transfer from area is missing for It is also the only subcategories (i.e. and current 2008–2011. There ministry of culture Sarajevo. country with orads, sidewals, expenditures, are considerable and other) dissagregated construction of etc. NOTE: There changes in data on assets and X). Statistical data is a sheet with revenues liabilities. contains area and all indicators catagories after also number of calculated (It 2009, and after staff. seems very well 2010. Interesting: 78 done: i.e per have number capita capital of employers, expenditure, and labor costs creditworthiness, per employer etc). Population dissagregated South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans dissagregated from census 2002, can calculate density with area and pop. Current expending YES YES YES YES YES Migh be able to YES YES Local calculate it Expenditures – Classification A Capital expending YES YES YES NO YES Might be able to YES NO calculate it Labor Costs YES NO YES NO YES YES YES NO Grants YES NO YES NO YES YES YN (Some Transfers NO Local might be actually Expenditures grants) – Classification B by economic Transfers NO NO NO NO YES YES YES NO categories Loan Repayment YES (Interest) NO YES (Combined NO YES YES NO (Interest and with other Fees) principal) (continued on next page) (continued) B&H Country Albania Albania FBiH RS Croatia Kosovo Macedonia Montenegro Serbia Number of local 80 but only 53 with 573 (2011), 576 government units reviewed data only (2010), 576 (2009), 170* (More with providing data 385 386 for 2010 63 576 (2008) 38 85 21 different regions) Years covered for analysis 2008–2011 2008–2012 2010 2008–2010* 2008–2011 2008–2011 2008–2011 2007*-2011 2008–2011** General public YES YES NO YES YES YN (Could be YES NO services a sum of other general categories) Defense NO YES NO YES YN (MBROJTJA NO NO (N/A) NO – PROTECTION?) Public order and NO YES NO YES YES NO YES NO safety Economic affairs YES YES NO YES YES YES YES YN (Road category) (sub-category (Transportation transportation) separate) Environmental YN (Sanitation with NO NO YES YES YES (Sewerage in YES YN (Water and protection (sc, Water supply) other category) Sewerage in same Local waste and wwater category) Expenditures – mangment) Classification B 79 Housing and YES (Housing and YES NO YES YES YN (Aggregated YES NO by Functional community water supply with with sewerage, categories ammenties (sc. sanitation and street lighting, etc) Housing water separate) supply) Health care YES YES NO YES YES YES NO (N/A) NO Recreation YN (Sports under YES (Aggregated with NO YES (Disagregated) YES (Aggregated YES (Aggregated YES (Aggregated NO Culture) Culture and Religion) with Culture and with Culture) with Culture) Religion) Culture and religion YES (Under Culture YES (Aggregated with NO YES (Disagregated) YES (Aggregated YES (Aggregated YES (Aggregated NO and Sport) Recreation) with Recreation) with recreation) with recreation) Education (sc, pre, YES (Aggregated) YES NO YES YES YES YES YES (Under schools primary secondary) category) Social protecion YES YES NO YES YES YES YES NO Own source YES YES but some YES YES YES YES YES NO inconsistencies reported Shared taxes NO YES but some NO NO NO NO YES NO inconsistencies reported Municipal Revenues Grants, Transfers and YES YES YES YES YES YES YES NO – General Subsidies Annex 2: Data Gap Analysis and Reclassification Local government YES NO YES YES* (Check if NO YES YES NO borrowing corresponds) (continued on next page) (continued) B&H Country Albania Albania FBiH RS Croatia Kosovo Macedonia Montenegro Serbia Number of local 80 but only 53 with 573 (2011), 576 government units reviewed data only (2010), 576 (2009), 170* (More with providing data 385 386 for 2010 63 576 (2008) 38 85 21 different regions) Years covered for analysis 2008–2011 2008–2012 2010 2008–2010* 2008–2011 2008–2011 2008–2011 2007*-2011 2008–2011** Tax revenues YES YES but Mixed YES YES YES YES YES NO (Dissagregated) (Dissagregated) (Dissagregated) Fees YES (Mixed) YES but Mixed YES but Mixed YES (Aggregated YES YES YES NO with Charges) (Dissagregated) Charges YES (Mixed) NO NO YES (Aggregated YES NO NO NO Municipal with Fees) Revenues – Own source revenues Rents NO NO NO NO YES NO YES (Named as NO Fees) Fines NO NO NO NO YES NO YES (Mixed before NO 2010) Own municipal NO YES YES YES (Property NO YES NO NO capital revenues Income?) National population NO NO NO NO NO YES NO NO Statistical data 80 – national National GDP YES YES NO YES YES YES YES (Up to 2010) YES in report Annual inflation NO YES NO YES NO YES NO NO Population YES YES YES except for YES YES YES YES YES capital South East Europe Municipal Finance Review – Local Government Finance in the Western Balkans type LGU Municipality (B) or YES (type of city and YES M-municipality, NO NO Class, from 1–6 Classification by Commune (K) or canton variables) (Administratiev C-city, CO-county depending on regions Quark (Q) status:capital=1, population other cities=2, municipalities=3) Statistical data – disaggregated GDP (unit) NO YES NO NO NO NO NO NO Density NO YES NO (but can YES NO (but can NO (but can YES NO (but can calculate it from calculate it from calculate it from calculate it from population and population and population and population and area) area) area) area) Rural/Urban NO NO NO NO (Maybe NO YES NO NO through LGU Cat) OECD Profile Tax revenues YES NO YES YES YES YES YES NO Revenues (Dissagregated) (Dissagregated) (Dissagregated) (Dissagregated) (Dissagregated) Non-tax revenues YN (Can calculate NO YES YN (Can calculate YN (Can calculate YN (Can calculate YN (Can calculate NO from difference) from difference) from difference) from difference) from difference) Grants YES NO YES YES YES YES YES NO (continued on next page) (continued) B&H Country Albania Albania FBiH RS Croatia Kosovo Macedonia Montenegro Serbia Number of local 80 but only 53 with 573 (2011), 576 government units reviewed data only (2010), 576 (2009), 170* (More with providing data 385 386 for 2010 63 576 (2008) 38 85 21 different regions) Years covered for analysis 2008–2011 2008–2012 2010 2008–2010* 2008–2011 2008–2011 2008–2011 2007*-2011 2008–2011** General public YES YES NO YES YES YN (Could be YES NO services a sum of other general categories) Defence NO YES NO YES YN (MBROJTJA NO NO (N/A) NO – PROTECTION?) Public order and NO YES NO YES YES NO YES NO safety Education YES YES NO YES YES YES YES YES Health YES YES NO YES YES YES NO (N/A) NO Social security and YES YES NO YES YES YES YES NO welfare housing and YES YES NO YES YES YN (Aggregated YES YN (Water and community with sewerage, Sewerage in same OECD amenities street lighting, etc) category) 81 Expenditure by functions recreational cultural YES YES NO YES YES YES YES NO and religious fuel and energy NO YES NO YES NO NO NO NO agriculture forestry NO YES NO YES NO NO NO NO fishing and hunting mining NO YES NO YES NO NO NO NO manufacturing and construction transportation and NY Transportation YES NO YES NO NO NO NY (Roads communitcation Category) other economic YES YES NO YES YES YES (Aggregated YES NO affairs with communal activities) other functions (–) YES NO (–) (–) (–) (–) (–) Annex 2: Data Gap Analysis and Reclassification Annex 3: Collection of Local Government Finance Data The MFR, under the overall capacity building frame- kindergartens, primary education (1–8 work of UPP, has supported the Local Government grades), secondary education (9–12) Associations in developing disaggregated municipal 4. Total expenditures on own/devolved and finance datasets. The MFR local experts in coopera- delegated functions tion with the partner LGAs collected fiscal reports by municipalities and composed some aggregate, b.  Local government revenues country wide public finance and economic data. The 1. Total own source revenues targeted period of data collection was 2008–2011. 2. Current own source revenues In this report, the most recent years (2010 or 2011) 3. Of this: local taxes are used for describing the present situation and 4. Capital own source revenues changes are assessed for the past four years, when 5. Total shared revenues data were available in most the participating entities. 6. Shared revenues allocated by origin (not by formula) Common requirements on the municipal finance 7. Total and main types of transfers (grants, sub- data were that only publicly available, official sidies, contributions, etc.): general, conditional and authorized data would be incorporated in (block, earmarked, equalization funds, etc.) the datasets. The terms and definition used by the 8. Capital investment grants municipal finance datasets were specified in the country-specific reports’ methodological annexes. c.  Municipal borrowing The municipal finance datasets, together with their 1. Local government borrowing: net; total, short explanatory notes and definitions, were made avail- and long term, main types (bank loan, bonds, able in English and local language. other) 2. Local government debt (stock) The disaggregated datasets followed the same, 3. Local government borrowing capacity (mea- standard structure with the following types of sured by the regulatory limit) information: 4. Local government debt repayment (interest and principal) a.  Local government expenditures 1. Local government current expenditures by These municipal finance data were supplemented d.  functions (sectors) and total with indicators on local government typology and 2. Capital spending by local service areas (func- statistical data. They help the analysis by allowing tions) and total various grouping of the disaggregated data and 3. Total current and capital expenditures by the making possible to create some indicators: main spending areas/services: water man- 1. Capital city, main cities, other municipalities agement (including water provision, waste 2. Urban/rural or any other special status water collection/treatment), municipal solid 3. Intermediary tier: county, planning/develop- waste management, public transportation, ment region 82 Annex 3: Collection of Local Government Finance Data 4. Population: total, main age groups 8. General government debt; 5. Area (km2) 9. Annual inflation rate 6. Other service performance indicators 10. Annual average exchange rate The general economic data aimed to support pri- e.  For the regional comparison, the national munici- f.  marily the international comparison in the region: pal finance datasets were reviewed in detail 1. Country GDP and revised, where necessary. Different report- 2. Total national budget expenditures and ing standards and practices across the region revenues; required reclassification and re-coding of data 3. Total public expenditures; during this process of data cleaning. For the data 4. Total national tax revenue; analysis, all values were adjusted for inflation 5. Total national tax revenue shared by origin; and correspond to 2008 prices. The data clean-up 6. Total public/national government assets; and reclassification followed the methodology as 7. Total national capital investments (or general described in preceding Annex 2. government (public sector) capital formation) SUMMARY OF MFR DATASETS (AS COLLECTED) Number of local government Years with available units providing data data Major gaps in the dataset Albania 385 2002–2011 (362 for 2002–2007) FBiH 57–70, depending on survey 2010 Survey based data collection with 88% response rate. tables Aggregate data are available 2008–2010 BiH RS 63 2008–2011 Revenues before 2010 and after are not comparable. Expenditures by functions are missing Croatia 573 2008–2011 The only country with data on assets and liabilities Kosovo 38 2008–2011 No information on borrowing Macedonia 85 2008–2011 Montenegro 21 2007–2011 Only a sample of expenditures by functions Serbia 169 2008–2010 Statistical Office data, not detailed. Expenditures by functions are not reliable, because include cen- tral spending at local level. Aggregate data are available from public sources (MoF website) 83