Bosnia and Herzegovina: Local Government Debt

Abstract

Fiscal decentralisation in Bosnia and Herzegovina is characterised by a complex post-conflict constitutional organisation consisting of two entities and one self-governing unit. The decentralisation process and the position of local government units in both entities over the past 20 years can be divided into three phases. Although the entities are constitutionally organised differently, the assignment of expenditures to local government units is similar in each one and did not change over time. At the same time, the allocation of revenues to local government units has improved significantly. However, since indirect tax revenues take the highest share of their total revenues in both entities, with the outbreak of global financial crisis these revenues became volatile, which triggered a rise in local public debts in both entities.

Introduction

The aim of this chapter is to analyse intergovernmental fiscal relations in an asymmetric confederative country: Bosnia and Herzegovina (BiH). We evaluate and assess the changes in the status of local government units (LGUs) in BiH over the last 20 years in terms of the assignment of expenditures and revenues. As BiH consists of two separate entities, as well as one self-governing unit, we aim to determine the position of LGUs in each of the two entities, since one is more fiscally decentralised than the other. A special focus will be placed upon aspects of the political economy of fiscal decentralisation in BiH, due to its unique constitutional organisation, which is reflected in the political and institutional dynamics that affect the speed of fiscal reforms in BiH.

The chapter is organised as follows: after the Introduction, we briefly analyse the constitutional and fiscal structure of BiH and fiscal reforms undertaken since 1995. Through a detailed analysis of fiscal reforms in BiH over the past 20 years, we have determined three phases that affected the position of LGUs’ revenue sources: the first phase from 1995 to 2006, the second from 2006 up to the outbreak of the global financial crisis (GFC) in 2008, and the final, third, phase from 2008 up to the present day. Our analysis shows that the share of LGU revenues and expenditures to total revenues/expenditures/gross domestic product (GDP) has been increasing over the past decade, which might indicate the successful implementation of the ideas presented in the first generation of fiscal federalism (Oates, 1972). However, the case of BiH has one peculiarity in comparison to other fiscally decentralised countries. The usual question about assessing the level of fiscal decentralisation is related to identifying the motivation of a central government to give up powers and resources to local governments (Eaton, Kaiser, & Smoke, 2010). Hence, politically motivated strategies conducted at the central level influence the size and sequence of political, administrative, and fiscal decentralisation (Eaton et al., 2010, p. 10). However, in BiH, we can determine that the opposite process occurred with the centralisation of indirect taxes at the state level during the second phase, which, in turn, brought multiple benefits to all subnational levels of government, including LGUs. A significant increase in the amount of collected revenues from indirect taxes brought about easier and more transparent LGU budget planning on one side, but caused a higher dependence on indirect tax revenues on the other. So, when a short-term fall in revenues from indirect taxes during the first years of the GFC triggered fiscal stress at the LGU level in both entities, LGUs began to run budget deficits and to borrow more. Even though LGUs in both entities were not assigned greater expenditure responsibilities during the third phase, they were affected by the fall in tax revenues and the central government’s fiscal consolidation and austerity plans. This is especially highlighted by the fall in LGUs’ capital investments and problems of increased local public debts. Furthermore, a status quo in the expenditure assignment at the level of LGUs might also be indicative of inefficient public spending (Antić, 2013). Large and growing local public debts are possibly the single most significant concern in terms of a sustainable provision of local public goods in the long run in BiH. The idea of maintaining hard budget constraints as one of five necessary conditions¹ defined in the second generation of fiscal federalism (Weingast, 2007) might seem appropriate now. We evaluate an option that the budget deficits that LGUs were running are not a consequence of structural vertical imbalances, but instead of spending decisions at the local level that increased the expenditure side of the budgets. These spending decisions have been brought in at the local level, which might indicate that local mayors are acting in their own-self-interest rather than in the interest of the community. This type of behaviour has also been recognised in the second generation of fiscal federalism (Weingast, 2007).

Territorial and Administrative Organisation of BiH since 1995

BiH became an independent state on 1 March 1992, but due to the war in BiH from 1992 to 1995, its constitutional and therefore fiscal structure could not be determined until the adoption of the new BiH Constitution (i.e. the Dayton Agreement²), which was signed in December 1995. Two structural elements of the constitutional fiscal federalism in BiH are as follows:

• The Constitution of BiH forms a federal structure, and although its text does not define decisively the type of country, it certainly shows a somewhat weaker structure for the Federation of BiH (FBiH)³; and

• From the economic point of view, the Constitution provides a solution of typical economic federalism, both in the full freedom of movement of goods, services, and capital within BiH. The authority is given to entities and institutions of BiH to constitute basic elements of an economic system (monetary system, customs, export trade, international financial relations, etc.) and key instruments of economic policy. (all [sic])

(Begić, 2000, p. 23)

The Dayton Agreement as an official BiH Constitution indicated the organisation of the country as an asymmetric confederation consisting of two entities: FBiH and Republika Srpska (RS). Brčko District⁴ (BD) was established through a final arbitration decision in March 1999 and is a self-governing administrative unit under the sovereignty of BiH. In both FBiH and RS, LGUs exist in the form of municipalities and cities, even though the Dayton Agreement does not define LGUs as a constitutional category (Mujakić, 2010, p. 1047). In FBiH, there is an intermediate level of government between the federal level and the municipalities—the Cantons. Hence, BiH is sometimes referred to as an asymmetric confederation, since the federal level (the FBiH entity) is below the state level (BiH). At the same time, FBiH is a fiscally decentralised entity (federal level, cantons, and municipalities and cities) whereas RS is a more centralised entity (RS itself, and its municipalities and cities). Each entity has its own Constitution, which must be fully compliant with the BiH Constitution (under Article 3.3). The current structure of BiH is shown in Fig. 6.1.

Fig. 6.1

Constitutional organisation of Bosnia and Herzegovina. Source: Kreso (2005, p. 256); Antić, (2013, p. 304). Note: Solid line indicates direct supervision; dashed line indicates indirect supervision. *Under Article 10 of the Dayton Agreement, OHR stands for “Office of the High Representative” which oversees the civilian implementation of the Dayton Agreement. The Principal Deputy High Representative serves as Brčko District Supervisor; since 1999, the number of municipalities in FB&H has decreased (from 84 to 79 due to new organisation of the City of Mostar). The City of Sarajevo in FB&H includes four municipalities in its structure, and the City of East Sarajevo in RS includes six municipalities. According to the data from RS Institute of Statistics (2016), there are 58 municipalities and 6 cities in RS

The idea for organisation of BiH in terms of a loosely fiscally decentralised country came from the previous organisation of Yugoslavia. In terms of significant hallmarks in the improvements of fiscal federalism in BiH, Antić (2014) identifies three phases which parallel those highlighted in this chapter. The first phase is characterised by the high level of asymmetry and extremely high degree of fiscal decentralisation. During this period, we cannot determine benefits of fiscal decentralisation at the level of LGUs in either entity since the entire country depended on significant post-war international aid for reconstruction and development. Hence, there is no national data available (for the period 1995–2003) to identify the position of LGUs since during most of the first phase all government levels were being established and operationalised. During this phase, two associations of local authorities were established in both entities (in 1998 in RS and in 2002 in FBiH) as LGU representatives. Their negotiating power increased steadily throughout three phases. In the first phase, LGUs were also assigned most local expenditures which changed insignificantly throughout three phases.

When we discuss the process of fiscal decentralisation in BiH and the possible success of fiscal federalism, we must bear in mind that state (BiH) government during most of first phase was very weak and that has caused most impediments to the speed and success of the entire decentralisation process in BiH. With the centralisation of indirect taxes in the second phase, the central government of BiH became functionally and institutionally more powerful, assignments for expenditures at the level of BiH increased, which all reflected in the integration of BiH’s economic space and removal of barriers to movement of goods and services, capital, and people. Additionally, Antić (2014) highlights the significance of the integration of BiH’s economic space as one of the necessary preconditions for a successful functioning of the second generation of fiscal federalism.

Under the BiH Constitution, monetary policy under an orthodox currency board arrangement is defined to be conducted at the level of BiH. This macroeconomic arrangement puts pressures and increases the significance of an adequate fiscal policy. With the improvements in the fiscal policy in the second phase, the third phase was characterised by the institutionalisation of fiscal coordination, the introduction of budget planning, and improvements in the legislation and administration of direct taxes at the levels of entities (Antić, 2014). However, this process is still unfinished, and this is reflected by the LGUs. Therefore, authors such as Antić (2013, 2014) have suggested further improvements to fiscal coordination in terms of an institutional framework for fiscal management, establishing relations between levels of government, and imposing obligatory rules for all governments. In terms of a political economy of fiscal federalism in BiH, Antić (2014) describes the relationship between entities as a form of competitive federalism with the aim of attracting foreign direct investment (FDI). She classifies some functions, such as inter-entity fiscal coordination, as a combination of cooperative and executive federalism (Antić, 2013, 2014).

In comparison to the pre-transition period, the current BiH territorial and administrative structure is characterised by a larger number of LGUs in the post-transition period (namely, there were 109 LGUs in 1989 (Savezni zavod za statistiku, 1989) in comparison to the current 143 LGUs). The growth in number of LGUs is—in the most part—a direct consequence of the peculiar post-war position of certain LGUs that were divided between the two entities. So, the “new” municipalities are organised along entity lines, with mostly weak administrative and fiscal capacities. This is contrary to any economic rationale, especially since the results of 2013 BiH Census indicate a significant fall in BiH’s total population to its current 3.5 million inhabitants.⁵ Currently, there are 2.2 million inhabitants living in FBiH, 1.2 million in RS, and 83,500 in BD (BHAS, 2016). The majority (52.3 per cent) of the BiH population lives in LGUs (municipalities and cities) that have between 10,000 and 50,000 inhabitants. Current laws on principles of local self-government in FBiH and RS⁶ define criteria for organisation of cities in terms of population (minimum of 30,000 inhabitants in FBiH and 50,000 in RS) or other criteria (e.g. a municipality with less than 30,000 inhabitants that is a cantonal seat in FBiH can become a city or, in RS, a municipality with less than 50,000 inhabitants, but with a defined status of developed LGU can become a city). The LGUs in RS mostly retained their administrative and territorial organisation before and after the transition process. Most administrative and territorial changes occurred in FBiH with the introduction of Cantons, which took on some of the LGUs’ traditional expenditure assignments. The reason lies in the fact that FBiH was constituted (in 1994) before the signing of the Dayton Agreement, and was, as such, incorporated into the current organisation of BiH. Early papers (Čaušević, 2001; Ivanić, 2000; Kreso, 2005) dealt with intergovernmental fiscal relations between higher levels of government, mostly omitting LGUs. Only international projects dealt with the position of LGUs (e.g. USAID, 2005) in terms of their position in the “new” revenue-sharing mechanism that was introduced with the implementation of value added tax (VAT) as of 2006.

Fiscal Decentralisation in BiH

The peculiar constitutional organisation of BiH causes several impediments to the implementation of tax policy conducted in BiH. Fiscal policy in BiH is divided, in terms of jurisdictions, between state and entity levels. Since 2006 indirect taxes (VAT, excise duties, customs, and road fees) are collected at the state level whereas direct taxes (personal income tax—PIT, corporate income tax—CIT , and property taxes) together with social security contributions (SSC) are regulated at entity levels (Lazović-Pita, 2015). During three phases of fiscal decentralisation, most changes occurred at the higher tiers of government, which affected LGUs and significantly improved their position in terms of the size of allocated revenues to LGUs.

The First Phase: Post-war Reconstruction, International Aid, and the Decentralisation of Revenues

As noted before, LGUs in both entities were not recognised in the Dayton Agreement (Mujakić, 2010) but were recognised in two entity constitutions and by the European Charter of Local Self-Government. As most of the first phase was characterised by international aid for reconstruction and development, budgets at all levels of government suffered from horizontal and vertical imbalances due to mismatch between the allocation of responsibilities and available funds. Hence, LGUs were heavily underfunded and the provision of public services was poor and dependent on international aid. However, this issue was identified as being in urgent need of rectification, and subsequently the Parliament of RS adopted the Law on Local Self-Government of the RS in 2004, which set out municipal competences in detail, and listed the sources of revenues, defining LGUs’ authority to set local charges and stamp duties. The first law related to the position of LGUs in FBiH was enacted in 1995. This law was subjected to several amendments prior to a new law being introduced in 2006 (Mujakić, 2010). For example, the legal provisions from 1995 had to be amended to include a definition of a city as an LGU (in 1996), in terms of the status of the city of Sarajevo (as a capital) and city of Mostar (in 1997). Since FBiH has additional sub-central level of government—canton unlike RS, during this phase, RS was more proactive than FBiH in its legal set-up regarding the position of LGUs. In RS, the role and authority of a mayor in LGUs was legally defined in 2005, whereas in FBiH this occurred in 2006. The law redefined political power locally and increased the stakes in the outcome of local governmental performance, due to the fact the mayor became responsible for developing and submitting municipal/city budget drafts for adoption by the municipal council. Before 2004, mayors were elected from the members of the municipal councils in both entities, but following the 2004 elections, mayors, together with members of municipal councils in both entities, have been elected in local democratic elections organised every four years. The only exceptions today exist in the cities of Sarajevo, East Sarajevo and Mostar, where mayors are still elected by the municipal councils.

During this phase, LGUs were assigned responsibilities over local expenditures which mostly remained unchanged throughout three phases (Table 6.2). In terms of revenues, Lazović-Pita and Štambuk (2015) give a detailed overview of tax reforms in both BiH entities per tax type, throughout three phases. In the early years of transition and during the first phase, revenues in FBiH were divided between the FBiH governmental level and the cantons, as defined in the law adopted in 1996.⁷ Revenues that belonged to FBiH level were customs, excise duties, and other federal stamp duties. Cantonal revenues were revenues from retail sales tax (RST), CIT , and schedular income tax (wage tax) and cantonal stamp duties. In 1996, LGU revenues were determined by cantonal regulations. LGUs had authority over communal taxes, fines, and so on and regulation over property taxes. The most buoyant taxes, RST and wage tax, were legislated and collected at the FBiH level by the tax administration, but the distribution of those revenues was regulated by cantons. Since 2001, when the FBiH law was amended, revenues from certain groups of excise duties also began to be shared between FBiH level and cantonal level. From 2003, RST was distributed between cantons and LGUs on a derivate basis. As a result, this created a situation in which LGUs enjoy different positions in each of the ten cantons. The share of revenues assigned to LGUs in each canton varied significantly and was dependent on cantonal or entity laws for each tax type. For example, Canton Sarajevo financed its LGUs by means of grants and shared property taxes. In 2003, total LGU revenues in FBiH were €175.32 m, while in 2006 total revenues were €286.31 m, giving an average annual increase of 17.7 per cent.

Prior to 2006, RST was shared based on cantonal laws, and cantons assigned varying proportions of RST revenues to the LGUs in their jurisdiction. Revenues from RST among ten cantons were mostly shared so that 75–80 per cent of RST revenues belonged to the cantonal budget, and 20 per cent or 25 per cent to LGUs’ budgets. The position of LGUs varied significantly in each canton—from Canton Sarajevo,⁸ which shared only property tax with its LGUs and otherwise financed its LGUs by allocating unconditional grants—to Tuzla Canton which shared most of its revenues with LGUs. This can be said for revenues from all taxes—wage tax, property taxes, and so on except the revenues from CIT which were shared between FBiH and cantons. USAID (2005) provided a detailed summary of the revenue-sharing mechanism per tax type in FBiH in the first phase.

In RS, the redistribution of revenues from RST was organised so that at least 30 per cent of collected RST revenues in the LGUs were redistributed back to LGUs, while the City of Banja Luka was entitled to 35 per cent of collected RST revenues. The other most important shared tax revenues came from the schedular wage tax and other property taxes, whereas CIT revenues were fully retained within the budget of RS. The Law on the Budget System of RS from 2003 specified that LGUs were entitled to receive 25 per cent of the wage tax generated in their jurisdictions. Hence, all tax revenues in RS were shared, except for property taxes, which were legally paid at the entity level, although revenues from property taxes entirely belonged to LGUs. During this phase, the total amount of LGU revenues in 2003 was €152.47 m, while in 2006 total LGU revenues amounted to €249.79 m, giving an average annual increase of 17 per cent which is similar to that of FBiH.

The 2003 Law on the Budget System of RS⁹ attempted to fiscally equalise LGUs. The Law divided municipalities into four categories, depending on the development index, and based on criteria adopted by the National Assembly of RS. According to the adopted criteria, developed municipalities received 30 per cent of RST revenues, semi-developed municipalities 40 per cent, underdeveloped 50 per cent and extremely underdeveloped 60 per cent of collected RST revenues in LGUs. Cities received 35 per cent of RST revenues. However, this system failed to equalise between the developed and underdeveloped LGUs, as assigning an additional percentage of RST revenues to underdeveloped LGUs with low fiscal capacities did not produce equalisation effects, and so the system was abolished.

During the first phase, the main characteristic in both entities was a very low level of autonomy of LGUs over revenues. The greatest proportion of revenues was transferred to LGUs in the form of unconditional grants. LGUs could not set their own tax base or rate for any major revenue source. Own sources, or non-tax revenues of LGUs, mainly came from different user fees and charges, various communal charges, income from renting, or building fees. In this phase, LGUs were not the focus of reform activities, as state-building was the priority for both the international community and local actors.

Long-term¹⁰ data availability for LGUs in BiH is scarce, and currently relies on aggregated datasets from the International Monetary Fund (IMF) Government Finance Statistics (GFS). Aggregated data is useful for cross-country comparison, but inevitably leads to a loss of information and a lack of possibility for pinpointing clear distinctions between own revenues and shared revenues, especially at the level of LGUs. Similarly, on the expenditure side we can only use IMF GFS classification of expenditures since functional classification of expenditures in BiH according to IMF GFS is available starting from 2013.

BiH collects significant amount of revenues, especially tax revenues measured as a percentage of GDP. The share of total consolidated government revenues¹¹ to GDP during the three phases was on average 43.3 per cent of GDP (2003–2016), and the share of tax revenues to GDP was on average 23.1 per cent. From Fig. 6.2, we can determine that LGU revenues in RS have an average share of 17.0 per cent of total revenues and an average share of 6.7 per cent of GDP (2003–2016). On the other side, due to another intergovernmental level—the cantons in FBiH—the share of LGU revenues to total revenues and GDP in FBiH is smaller, amounting on average to 10.2 per cent and 4 per cent respectively (2003–2016). The share of LGU revenues in both entities in absolute amounts is nearly identical on an annual basis. During this phase, the position of LGUs improved significantly in terms of the size of allocated revenues even though we have data available as of 2003. During this period, LGUs were not assigned greater expenditures nor were given the authority over any additional revenues.

Fig. 6.2

The share of LGUs’ revenues to total revenues and to GDP in RS and FBiH. Source: CBBiH, 2016,Footnote

All data are taken from the Central Bank of BiH and is in accordance with the IMF GFS (2001) methodology. Therefore, total revenues include entities’ general government revenues in accordance with the definitions from IMF GFS (including extra-budgetary funds). Authors would like to thank the CBBiH statistics department and government finance statistics staff of the CBBiH for their valuable insights and explanations regarding the methodological issues.

own interpretation. Note: LR—local revenues; TR—total revenues

The Second Phase: Consolidation of Indirect Taxes in BiH and the Decentralisation of Revenues

In 2006, the Parliament of the Federation of Bosnia and Herzegovina adopted the Law on Principles of Local Self-Government in FBiH, which redefined the role of LGUs. Unlike in the case of RS, where the law was explicit in the determination of LGUs’ responsibilities, in FBiH the law was general, and indicated that each canton has the authority to determine the responsibilities of LGUs within the canton itself. To date, most cantons have adopted LGU-related laws, even though some cantonal laws in this respect have still not been brought. The most significant change from the previous system has been the legally defined ability and formal regulation of LGUs regarding the borrowing and/or issuing of bonds. In terms of responsibilities of mayors in FBiH and RS, Mujakić (2010, 2012) provides a detailed analysis. His research shows that with the legal changes in both entities, mayors’ relationships to municipal councils changed, and that mayors gained greater responsibilities. Surveys conducted among citizens have shown that the local level of government enjoys the highest degree of trust from citizens in comparison to other tiers of government (Analitika, 2014; Transparency International, 2014). The relationship between council and directly elected mayor brought new positive dynamics at local level, in the sense that accountability was strengthened through mayors’ sole responsibility for quality of services at local level. The mayor could not be recalled by the municipal assembly any more, but only by the referendum of all registered voters in municipality. Mujakić (2010) also analysed the revenue position of the LGUs in FBiH and RS (without an analysis of expenditure assignment). His research indicated that LGUs in RS were—in terms of revenue allocation—in a better position, receiving a greater share of total revenues than LGUs in FBiH. However, measured by the Organisation for Economic Cooperation and Development (OECD) methodology and internationally comparable indicators, Antić (2013, p. 291) indicates that fiscal autonomy of local communities in BiH is “very low considering that the share of revenue on which they decide entirely (non-tax revenue) or partially (tax on property) is low”.

We have already determined that the preconditions for implementation of the second phase began in the first phase. In 2003, the High Representative for Bosnia and Herzegovina proclaimed a unified Law on Customs and Excise Duties, which provided a basis for further developments in indirect taxation and the proclamation of the Decision on the Establishment of the Commission for Indirect Taxation Policy.¹³ Since customs policy is the only constitutional provision regarding fiscal policy harmonisation in BiH, customs administrations were incorporated into an independent institution, the Indirect Tax Authority (ITA), which is the institution responsible for the collection of all indirect taxes in BiH (i.e. excise duties, VAT, customs, and road excise). The ITA became operational in 2005 when the VAT law was adopted. VAT was introduced in 2006, at a single rate of 17 per cent. With the centralisation of indirect taxes at the state level, the sources of financing for BiH institutions changed considerably (Fig. 6.3) since up to that point, entities contributed towards BiH budget (two thirds of the budget were financed from FBiH and a remaining third by RS).

Fig. 6.3

Distribution of indirect tax revenues in BiH. Source: Antić (2013, p. 286)

Inter-entity and BD revenue-sharing mechanism is determined (mostly) on an annual basis by a decision from the ITA board of directors,¹⁴ so room for improvement in this area still exists. Hence, the percentage of indirect tax revenues belonging to each entity varied from 2006, and, at present 63.83 per cent belongs to FBiH, 32.62 per cent to RS, and 3.55 per cent to BD (ITA, 2016). Figure 6.3 shows the current division of indirect tax revenues between BiH and the entities, whereas inter-entity distribution of indirect tax revenues is regulated by two independent laws (one for each entity). These laws also define redistribution of other sources of revenue, namely direct taxes. This entire revenue-sharing mechanism is relevant for evaluation of the overall process of fiscal decentralisation in BiH since it shows that the political economy behind the decentralisation process and the operationalisation of fiscal federalism in BiH is determined by setting clear rules of intergovernmental fiscal relations at higher tiers of government and not at the level of LGUs. This means that as indirect tax system was clearly established, LGUs in both entities were in a better position in comparison to prior times especially in terms of budget planning as well as the size of revenues belonging to LGUs (Figs. 6.4 and 6.5).

Fig. 6.4

FBiH LGUs’ structure of revenues 2003–2016, in millions of EUR. Source: Central Bank of Bosnia and Herzegovina (2017), own calculation

Fig. 6.5

RS LGUs’ structure of revenues 2003–2016, in millions of EUR. Source: CBBH (2017), own calculation

In FBiH, based on the Law on Allocation of Public Revenues from 2006, the vertical division of revenues from indirect taxes is defined as in Fig. 6.3, so 8.42 per cent is being allocated to municipality/city level. The horizontal allocation is based on several weightings that capture the expenditure needs of jurisdictions and their fiscal capacities (Table 6.1).

Table 6.1 Criteria for revenue sharing of indirect taxes in FBiH

In addition to this, the formula contained weighting for the special spending needs of Canton Sarajevo, which is to a factor of two and three other cantons with the smallest revenue from RST in the derivation-based system, which were also given a special weighting (1.1, 1.5, and 1.8). A newer formula also incorporated a weighting factor of 1.2 for LGUs with over 60,000 inhabitants, and 1.5 for LGUs, which cover the costs of their elementary schools.¹⁵ The Index of Development measured the fiscal capacity factors of the LGUs. It was calculated as the average of the collected revenues from RST and Wage Tax for respective LGUs for 2005. Weightings for fiscal capacity were in the range of between 1.8 for LGUs with below 20 per cent of the FBiH average collection of taxes and 1.2 for LGUs with below 80 per cent of the average. At least 28.5 per cent of Wage Tax (replaced by PIT in 2009) revenues collected within the jurisdiction of LGUs was also assigned to LGUs in FBiH. The new system had a phase-in period of six years (2006–2011), over which time the new municipal share was increased. Exceptions were LGUs in Canton Sarajevo which continued to receive unconditional grants from the cantonal government until 2014. The allocation of CIT and property tax revenues remained the same as in the first phase.

One of the greatest disputes in the implementation of the 2006 Law on Allocation of Public Revenues was the unreliable data on population. BiH was not able to carry out a census of its population until 2013. The census prior to this was undertaken in 1991, the last pre-war year. The question of a new census was highly politicised, and continues to be a debatable issue, although official results have been published. For some LGUs, the 2013 Census results revealed a significant underestimation of population, which had direct implications on the calculations of LGUs’ share in horizontal distribution of, primarily, indirect tax revenues. Other municipalities/cities saw increases in their revenues from indirect taxes. However, there has been no significant change in the indirect tax revenue sharing in either entity since 2006 or in the third phase. In FBiH, it is expected that the new law on allocation of revenues will consider the results from 2013 Census, which will certainly reflect on small LGUs.

The Law on the Budget System of RS was amended in 2006, so that indirect tax revenues could be distributed vertically and horizontally. Initially, the vertical share from indirect tax revenues was as follows:

• 73.5 per cent, RS government

• 23.0 per cent, LGU level

• 3.5 per cent, Road Fund

The criteria for horizontal distribution of indirect tax revenues between cities and municipalities in 2006 were as follows:

• 75 per cent based on population

• 5 per cent based on area

• 15 per cent based on secondary school pupils

• 5 per cent based on elementary school pupils

Additionally, in 2006 special weighting for horizontal distribution of indirect tax revenues was assigned to the City of Banja Luka and several smaller LGUs. The phase-in period of the new system was six years. The RS government amended the law in the same year, increasing the vertical share for LGUs to 24 per cent and decreasing its own share to 72.5 per cent of indirect tax revenues (Fig. 6.3), which remained to date. More importantly, the government of the RS extended the phase-in period to ten years, and changed the criteria for horizontal distribution:

• 75 per cent based on population

• 15 per cent based on area

• 10 per cent based on secondary school pupils

As before, Wage Tax was shared 75 per cent for the Entity budget, and 25 per cent for LGUs’ budgets. CIT revenues were retained at the entity level, and other revenues, such as different concessions for natural resources, were also shared with LGUs. The allocation of CIT and property tax revenues remained the same as during the first phase. With the introduction of a new revenue-sharing system, the provision from the Law on the Budget System of RS from 2003 on equalisation between municipalities ceased to exist, and weights attached in the previous amendment to some underdeveloped municipalities were also removed.

Throughout this phase, LGUs in both entities received a higher share of indirect tax revenues because of the new revenue-sharing mechanism. There was a significant increase in revenues from property taxes and wage tax as an expected result of overall stronger economic growth in both entities. During this phase, LGUs were not given the authority over any additional revenues.

The Third Phase: The Outbreak of GFC and Rising Public Debt

After initial growth of revenues from indirect taxes (2006–2008), the global crisis affected the budgets and triggered fiscal stress across all levels of government. During the third phase, LGUs were affected by the fall in tax revenues and the central government’s fiscal consolidation and austerity plans, which was reflected in the running of budget deficits by LGUs, and their beginning to borrow more.

Regarding this phase, in terms of tax reforms, both entities went through PIT reforms in 2009, switching from schedular to synthetic PIT , with a flat tax rate of 10 per cent applied in FBiH. Furthermore, with the changes of PIT from schedular to synthetic, in FBiH, cantons had to harmonise tax laws under their jurisdiction. In RS, PIT reform included introduction of a flat PIT rate at 8 per cent, but in 2011 this was increased to 10 per cent. Throughout this phase, RS’s PIT law was changed and amended numerous times, affecting the share of PIT to total revenues (and the PIT share belonging to LGUs). Simultaneously, CIT went through several legislative changes, but in the third phase the CIT rate was set at 10 per cent in both entities. We cannot say that during this phase the tax reforms were deliberately aimed at centralisation of revenues since they were planned prior to the outbreak of the GFC. The centralisation of revenues brought greater fiscal discipline, simplified the tax procedures, increased simplification of the tax system, and the elimination of myriad (cantonal) tax codes which, in turn, brought benefits to LGUs in terms of greater revenues (Figs. 6.4 and 6.5).

Since 2008 and the outbreak of GFC, the redistribution of indirect taxes in both entities has remained virtually unchanged. The only (minor) alteration appeared in 2014, when LGUs in Canton Sarajevo were directly included in revenue sharing of indirect taxes from the entity level, replacing the unconditional grants allocated to them by the Canton. After the City of Sarajevo appealed before the Constitutional Court of FBiH because it was not included in revenue-sharing mechanisms, as is the case with other LGUs, the case was positively resolved in 2015.¹⁶ Association of local authorities in FBiH played a significant role in this process indicating the increasing negotiating power of LGUs in FBiH.

In FBiH, with the adoption of synthetic PIT from 2009, the minimum share of LGUs in PIT was increased to 34.46 per cent, and LGUs in Canton Sarajevo were given 1.79 per cent of PIT revenues. The Law on Allocation of Public Revenues in FBiH also provided for cantons to further allocate to LGUs some revenues that had been assigned exclusively to cantons, such as CIT revenues. Revenues from property taxes continue to exist as shared revenues between cantons and LGUs. These legal changes have put LGUs in FBiH in a significantly better position compared to prior periods.

In 2008, the RS government amended the Law so that the vertical share of the Road Fund was increased to 4 per cent, LGUs would receive an additional percentage point from indirect tax revenues (from 23 per cent to 24 per cent, Fig. 6.3) and the share of entity budget was decreased to 72 per cent. With the new Law from 2012, the shares and main revenues to be distributed remained the same, namely real estate tax¹⁷ remained 100 per cent allocated to LGUs, in addition to different fees and charges that are regulated by municipality and city councils. Real estate tax is regulated at the level of RS, but the law also defines LGUs’ authority in determination of market-based tax base (through zoning) and tax rates (which cannot be higher than 0.2 per cent). So, during this phase, LGUs in RS were also put in a better position in comparison to prior periods.

If we observe the share of LGUs’ revenues by tax type (Figs. 6.4 and 6.5) in both entities, we can determine that the highest share of all tax revenues is that of indirect tax revenues, followed by PIT revenues. Shared taxes in FBiH (income tax, VAT , and property tax) account for the largest share of LGUs’ tax revenues in FBiH. Revenues from taxes on income, profit, and capital gains¹⁸ throughout the 14-year period were on average 6 per cent of LGUs’ total revenues. Revenues from taxes on goods and services to LGUs’ total revenues amounted on average to 27.3 per cent in 2003–2016 and have recorded the largest increase amongst all shared taxes (from 20 per cent in 2003 to 30 per cent in 2016). In the period of the GFC, the revenues of LGUs began to decline. LGUs compensated this decline with an increase in non-tax revenues (in a broad sense, own revenues, Fig. 6.4). The share of other revenues to total LGU revenues (mostly coming from sales of goods and services) have been steadily increasing, and, on average, amounted to 29 per cent in the period from 2003 to 2016, but with more than 40 per cent share of total LGU revenues in 2016. Revenues from property income to total LGU revenues amounted on average to 8.7 per cent (2003–2016) but have constantly declined in the third phase. Taxes on property, which are traditionally local revenues, amounted on average (2003–2016) to 10.4 per cent of LGUs’ total revenues. Like revenues from property income, these revenues have been constantly falling, especially in the third phase, except for 2016. Grants (including both capital and current) account for, on average, 14 per cent of total LGU revenues (2003–2016). In 2015 and 2016, the share of grants to LGUs’ total revenues dropped to 3.5 per cent and 3.9 per cent respectively. The reason for this significant fall could be explained by the position of LGUs in Sarajevo Canton prior to 2015.

A similar situation to that of FBiH can be observed in RS. Revenues from tax on income, profit, and capital gains¹⁹ take on average (2003–2016) as little as 5.0 per cent of LGUs’ total revenues. However, these have been on the increase, especially after 2011. Revenues from taxes on goods and services are the most significant source of shared taxes and have, on average, amounted to 50.3 per cent of LGUs’ total revenues (2003–2016). Revenues from taxes on property take on average as little as 3.7 per cent (2003–2016) to LGUs’ total revenues. This is almost three times less than the same share in FBiH. Revenues from property income to LGUs’ total revenues amount to 2.4 per cent (2003–2016), but this proportion has exhibited extreme fluctuations across the three phases. Other revenues in RS (mostly from sales of goods and services) account for, on average, 27.8 per cent (2003–2016) of LGUs’ total revenues. Grants as share of total LGU revenues in RS show great variations, averaging around 5.3 per cent (2003–2016), but with peaks of 15 per cent and 13 per cent of total revenues in 2007 and 2009 respectively. In absolute amounts, in 2003, LGUs’ total revenues were €152 m, whereas in 2008 they had increased to €316 m, with the lowest annual total of €268 m being in 2014. LGUs in RS are much more dependent on VAT , which makes them more vulnerable to fluctuations in consumption (Fig. 6.5). This was evident in 2009 and 2010, as those years were marked by a sharp downturn in consumption, and hence VAT revenues. LGUs in RS were not able to fully compensate for these revenues, unlike in FBiH, by increasing own-source revenues.

Additionally, other revenues,²⁰ in addition to property income (other than Property Tax), can be assessed as own-source revenue. Through this type of revenue, LGUs can exercise some control, either by setting the rate or base, or varying collection effort. This can be said for LGUs in both entities. Property taxes include taxes on the use, ownership, or transfer of wealth (IMF GFS, 2001). In FBiH, revenues from property taxes are usually shared between cantons and LGUs, whereas revenue from real estate transfer taxes is—to the greatest extent—entirely LGUs’ own revenue. Revenues from real estate transfer tax account for the bulk of revenues from property taxes, peaking in 2008. The situation of property tax reforms in RS has already been explained. However, with the abolition of real estate transfer taxes and the adoption of a real estate tax since 2013 in RS, tax revenues have remained more-or-less stable, which might indicate the government’s intention to run a revenue-neutral tax reform.

If we look at LGUs’ total revenues in both entities, we can determine that at no point in the post-crisis period did revenues of LGUs drop below 2006 levels, that is, prior to the introduction of the new revenue-sharing system and VAT (Figs. 6.4 and 6.5). Cities and municipalities in both entities were not given any new expenditure assignments during this period of revenue surge, so even though some vertical imbalances existed in the period preceding VAT introduction, these were corrected with the introduction of the new system. As none of the LGUs previously set responsibilities were not abolished or limited, both entity laws on revenue distribution state that the revenue transferred to LGUs can be spent entirely according to budgets approved by local councils. These facts together with increasing allocation and the size of local revenues indicate that comparing to early years of decentralisation when LGUs were mostly neglected due to state-building process, their position improved significantly up to date. This might indicate slow but a successful implementation of the ideas presented in the first generation of fiscal federalism.

Assignment of Expenditures in Both Entities in Three Phases

Throughout the observed period, LGUs’ assignment of expenditures in both entities virtually did not change. Table 6.2 summarises LGUs’ assignment of expenditures in both entities throughout three phases. We can determine that local expenditures stayed more or less the same and that the only change occurred in the second phase, with the strengthening of the state (BiH) level when some expenditures were rightfully transferred to a higher governmental tier, such as defence.

Table 6.2 Changes over assignment of expenditures^a in Bosnia and Herzegovina, 1995–2016

The shares of LGU expenditures to total expenditures in both entities were very low, especially during the first phase. For example, the share of FBiH LGUs’ expenditures to FBiH total expenditures (2003–2005) amounted to 14.6 per cent and in RS 26.5 per cent in the same period. Since 2006, this share has increased in FBiH and amounted on average (2006–2016) to 16.6 per cent but remained the same in RS (average of 26.0 per cent, 2006–2016, but with significant drop in 2016, because of the integration of the extra-budgetary pension fund into the annual budget of the RS government, own calculation). This indicates that LGUs in FBiH were radically underfunded when compared to LGUs in RS, bearing in mind that LGUs’ assignment of expenditures in both entities is very similar (Table 6.2).

From 2006 to the present day, LGUs in both entities have had similar assigned functions, but with some notable distinctions. In the RS entity, LGUs are assigned the so-called material expenses of elementary and secondary schools, as well as pupil transport. In three cantons within FBiH, LGUs are given the function of financing material expenses of elementary schools and pupil transport. Also, in terms of social protection, LGUs in RS were, until 2012, almost exclusively assigned the provision of all types of social assistance, while in FBiH LGUs finance the functioning of local welfare centres, but social assistance is provided either at a cantonal or FBiH level. Furthermore, LGUs in RS are mandated by the Law on Social Protection of RS²¹ to earmark at least 10 per cent of their budget for social protection expenditures.

In FBiH, the share of compensation for employees and social contributions was, on average, 30.2 per cent of LGUs’ total expenditures, with a declining trend since 2013. Material expenses account for, on average, 19.1 per cent of LGUs’ total expenditures. Grants show strong variations, from 0 per cent of total expenditures to 8.5 per cent in 2011, but on average account for a small share of, on average, 3.3 per cent of LGUs’ total expenditures (2003–2016). Subsidies and interest on average account for 2.1 per cent of LGUs’ total expenditures, but other expenses account for, on average, a significant 22.6 per cent of LGUs’ total expenditures (Fig. 6.6). LGUs in FBiH saw strong growth of total expenditures after 2005. In 2008, total expenditures were €416 m, which means that total expenditure grew at an annual rate of 19 per cent. The largest share of growth was generated by expenditure on non-financial assets, peaking in 2008 at €67 m, revealing the fact that LGUs invested in capital projects from the increased revenues (Fig. 6.8). In 2009, total local expenses dropped sharply by 18 per cent, which reflects the GFC and associated decrease in total revenues. It is interesting to mention that in the period 2003–2008, the average annual growth of expenditure on employee remuneration was 11 per cent. After 2008, LGUs managed to adapt their expenditure by restraining the growth of this expenditure to a 1 per cent annual average but keeping expenditures for non-financial assets at approximately the same level nominally, albeit with sharp drops in 2009 and 2013.

Fig. 6.6

FBiH, LGUs’ structure of expenditures, in millions of EUR. Source: CBBH (2017), own calculation

Similar to cities and municipalities in FBiH, LGUs in RS increased their total expenditure from €154 m in 2003 to €365 m in 2008, at annual rate of 19 per cent, primarily to finance non-financial assets. However, in the period between 2009 and 2016, total expenses decreased at annual rate of 2 per cent, hitting an absolute minimum of €262 m in 2015 (Fig. 6.7). Expenses for employee remuneration increased until 2009 (although LGUs in FBiH had already lowered this expense by 2009) at an average annual rate of 15 per cent. Total expenses for Compensation for Employees and Social Contributions account for, on average, 28.6 per cent of LGUs’ total expenditures. In the first phase, these expenditures witnessed a decreasing trend, but have since (second and third phases) increased significantly. Similarly, Material Expenses had a downward trend until 2008 when this expense amounted to 19 per cent of LGUs’ total expenditures but have also been increasing since 2008. Other Expenses amounted, on average, to 11.8 per cent of total LGU expenditures, and interest and subsidies account for 1.9 per cent of LGUs’ total expenditures. LGUs in RS do not classify grants as expenditures. Looking at the capital grants of LGUs in RS, in 2008, the total expenditure for non-financial assets was €132 m and in 2009 it stood at €131 m (Fig. 6.8). Other expenses include property expenses other than interest (e.g. dividends and rent) and miscellaneous other expenses (IMF GFS, 2001). Due to declining revenues from 2008, because of a combination of the economic crisis and the entity’s foreign debt servicing issues, LGUs in RS decided to cut down on capital expenditure, but employee remuneration still increased at an annual rate of 3 per cent. The share of this expense to total expenditures also increased, because of decreasing revenue and consequentially the total expenditure of LGUs.

Fig. 6.7

RS, LGUs’ structure of expenditures, in millions of EUR. Source: CBBH (2017), own calculation

Fig. 6.8

Expenditure for non-financial assets of LGUs in BiH, in millions of EUR. Source: CBBH (2017), own calculation

It was not possible to obtain numbers of employees within local administrations in either entity, but it is reasonable to assume that the numbers have increased due to increased revenues in the period 2005–2008. After 2009, and a downturn in revenues, it became almost impossible to downsize the number of local authority employees to compensate for revenue losses. Antić (2013) analysed the position of LGUs in FBiH and RS in 2008 and showed that in FBiH the local level has increased wage expenditures at the expense of the social benefits whereas in RS LGUs have retained their wage expenditures at the same level and reduced grants, social transfers, and capital expenditures.

Expenditure for non-financial assets declined sharply in the third period, 2009–2016. Because of the structure of FBiH, capital investments are financed by the cantonal and FBiH governments, making the capital expenditure of LGUs in FBiH much smaller in comparison to LGUs in RS. Those expenditures dropped sharply after the GFC in 2008 and in subsequent years because of declining revenues from indirect taxes due to foreign debt servicing reducing the funds available for LGUs (Fig. 6.8).

We can determine that even though assignment of expenditures in both entities did not change throughout the observed period, the size of local expenditures changed during three periods. During the second phase and most of the third phase, local expenditures in both entities increased significantly especially in two categories: compensation of employees and material expenses. An increase in unproductive expenditures such as compensation of employees is always debatable and might indicate inefficient spending at the local level. However, parallel to this process LGUs in both entities increased capital investments especially in the second phase but as a fall in revenues triggered fiscal stress since 2009, they began to decline up to date.

Increasing Local Public Debts: Current Issues

With the outbreak of the GFC and inability of all governmental levels to tackle high public spending in relation to falling public revenues since 2009, BiH (and both of its entities) began to increase its external public debt, which is also valid for LGUs. Growth in revenue from indirect taxes in 2006 and 2007 launched a spiral of expenses of a complex administrative apparatus and social benefits at all levels of government. Obligations created under collective agreements and laws on social rights quickly melted fiscal surpluses from 2006 and 2007. Because of the rigidity of wages and social benefits that are regulated by Entity laws, governments were not able to respond quickly to the revenue decline caused by the economic crisis and implementation of the Stabilization and Association Agreement with the European Union (EU). Stand-by arrangement with the IMF from 2009 was only partially carried out, since the government had little power to enforce austerity plans due to political aims (Antić, 2013). Similar occurred with LGUs, which is evident from Fig. 6.9.

Fig. 6.9

Size and the structure of local government debt in all municipalities, 2003–2015 (millions of EUR). Source: CBBH (2017), own calculation

Figure 6.9 indicates a significant increase in the public borrowing of LGUs in BiH (across both entities) after 2009. LGUs heavily borrow from the banking sector in BiH, even though some have issued bonds. Due to a lack of data availability, figures do not include possible tax arrears of LGUs’ public companies. However, in the observed period (from 2008, when the first LGU securities were issued, to 2015), securities accounted for an average of 12 per cent of total borrowing. Therefore, the most pressing issue at present that both entities and LGUs face relates to increased borrowing by means of both commercial banks and the issuing of bonds, which imposes a threat to future capital investments (especially for LGUs). Furthermore, with greater public debt burden of all government levels, LGUs are in an especially difficult position about covering all public expenditures and efficiently providing local public goods.

LGUs’ borrowing capacity is legally determined by two entity laws²² regulating overall (consolidated) entity borrowing. Legally set limits could be assessed as a hard budget constraint as there are no legally defined bail out options. The FBiH law permits LGUs to borrow both domestically and from abroad, in local and foreign currency, but total debt is denominated in local currency (the same is regulated in RS). The FBiH law has been subjected to several amendments, but the most important legal provision deals with setting limits to LGUs’ long-term borrowing (loans and securities-mostly bonds). LGUs can generate long-term debt (debt servicing longer than a year) only under the condition that, at the time of the loan’s approval, debt service for each consecutive year, including servicing of the new loan and all loans for which an LGU has issued guarantee(s), does not exceed 10 per cent of revenues collected in the previous fiscal year. Short-term debt needs to be paid out within one fiscal year and cannot be refinanced or extended to beyond the end of the fiscal year. This debt cannot exceed 5 per cent of the LGU’s revenues in the previous fiscal year (Basarić, 2009). In the RS, the first law regulating debt was adopted in 2007, but this was changed in 2012. Currently LGUs in RS can generate long-term debt under similar conditions as in FBiH, but the limit is set to 18 per cent of revenues. Short-term debt limits for LGUs in RS are the same as in FBiH (5 per cent). However, inter-entity differences exist in terms of approval procedures from higher tiers of government, setting guarantees, and so on as well as the entity’s responsibility (or lack thereof) regarding an LGU’s inability to repay and manage its public debt. Additionally, two aforementioned entity laws regulating LGUs’ status in terms of the law on local self-government also define the borrowing procedure for LGUs. The greatest threat to the successful implementation of fiscal decentralisation at the level of LGUs in both BiH entities comes from the possible idea of an unleashing of LGUs’ increased borrowing through changes in entity’s laws that regulate local indebtedness. The idea of maintaining hard budget constraints on LGUs’ borrowing, stemming from the second generation of fiscal federalism, might seem appropriate at present times.

The BiH Ministry of Finance and Treasury has also recorded LGUs’ borrowing in terms of loans taken from commercial banks and securities issued since 2010. By their classification of domestic public debt of each entity, LGUs have different borrowing strategies: LGUs in RS rely on loans from commercial banks more heavily than LGUs in FBiH, which is also a result of the legally set limits. Table 6.3 shows the share of LGUs’ domestic debt to total domestic public debt in FBiH and RS. The difference between LGU borrowing in FBiH and RS is significant, as it comes from legal provisions within the two entities, and, in the given period (2010–2015), RS LGUs’ domestic public debt as a proportion of total RS domestic public debt was, on average, almost four times larger than the same share for FBiH LGUs.

Table 6.3 Share of LGUs’ total domestic debt to entity’s total domestic debt, 2010–2015

Even though it is perceived that, with clear legal provisions regarding LGUs’ borrowing limits, the system has become more transparent, this has only recently become a practice. In fact, LGUs’ public data are only recently available, so we cannot determine the final purpose of the increased borrowing in the long run. We would assume that loans have been taken out for capital investment purposes, but they might also be used for financing current deficits or LGUs’ public companies’ tax arrears. Hence, it is necessary to introduce an obligatory and transparent system of reporting for all LGUs so that it is possible to evaluate and compare the actual provision of local public goods in LGUs across both entities.

Conclusion

Since 1995, the LGUs in BiH have passed through three phases. In the period 1995–2004, LGUs mostly functioned in a manner based on the former Yugoslav local governance system, meaning that mayors were elected by local councils and the scope of functions of LGUs was unclear. The introduction of the direct election of mayors in 2004 was the first major reform in this sector. This change allowed for greater accountability at the local level and has made local administration more transparent in its work. Both entities adopted new legislation on local self-governance in this period, with this clearly defining the functions of LGUs, and making them independent in their scope of work. Such legislation allowed for the possibility of delegating some functions to LGUs whereas the assignment of revenues for performing those functions was regulated by different laws. As the state-building process helped by the international aid for reconstruction and development was the primary aim during most of first phase, LGUs’ needs and position were mostly neglected but began to improve with the new local legal provisions in both entities as of 2004. National data for most of the first phase (1995–2003) is non-existent, so our conclusions are drawn based upon the data from 2003 onwards.

LGUs’ position improved significantly during the second phase—since 2006, as a new set of revenue-sharing laws was enacted in both entities. This came to be a necessity after RST was replaced by VAT at the state level. The new laws meant that LGUs had, for the first time, a predictable, transparent, and stable revenue source. Due to the strong growth of revenues from indirect taxes in the second phase, cities and municipalities significantly increased their revenues right up to (and slightly beyond) the GFC in 2008. In terms of local expenditures, they increased as well especially the category of compensation of employees. It was not possible to obtain data on the number of employees in LGU administrations, but expenditure on employee remuneration almost doubled in the period 2003–2008, and then remained at approximately the same level until 2015.

In terms of LGU revenues as a share of GDP, strong growth was witnessed in the period 2003–2008 in both entities. From 2009, this ratio declined, but LGUs in RS have been impacted more negatively. In fact, measured by this indicator, the position of LGUs in RS has been comparably far worse than that of LGUs in FBiH. Even though fiscal autonomy of LGUs in both entities is very low in revenue terms, LGUs enjoy autonomy in spending their budget based on local council decisions, within the mandated functions of LGUs. However, most revenues of LGUs are collected by higher tiers of government, and distributed to LGUs based on formulae, and this is especially the case in RS, where more than 60 per cent (depending on year) of LGU revenues come from indirect taxes. LGUs’ spending decisions are not related to the level of effort they invest in collecting revenues, nor on deciding on tax base or rate to best suit local preferences, hence their fiscal autonomy is low. With the outbreak of the GFC, LGUs were inert in adapting the expenditure side, as they expected that upper tiers would distribute the same amount of revenues as before. However, higher tiers of government are not obliged to do so as there is no defined bail out options for LGUs in either entities. This caused budget deficits and issues related to increases in LGUs’ public debt or increased fiscal stress at the level of LGUs. Increasing local public debts are currently the single most significant concern in terms of a sustainable provision of local public goods in the long run in BiH. The repayment of loans is taking a toll on local revenues. LGUs in RS are more dependent on the sharing of indirect taxes, and therefore the negative impact of foreign debt service is greater for LGUs in this entity. Furthermore, LGU’s low fiscal autonomy and the fall in revenues allocated to LGUs in both entities affects their increased borrowing, both through loans (the majority) and the issuing of bonds.

Another issue is how well existing revenue-sharing laws capture the expenditure needs of LGUs. At present, there are no clear criteria for setting the expenditure standards for services delivered either by cantons or municipalities in either entity. In general, due to decentralisation of the administrative system, it has become increasingly difficult to perform coordination and oversight of the quality of services delivered. As LGUs are running deficits and as there is no criteria set for the evaluation of expenditures at the level of LGUs, locally elected representatives might act in their own self-interest. Hence, this might be a future threat to the successful implementation of fiscal federalism identified in the second generation of fiscal federalism.

The decentralisation process in the 1990s in BiH was primarily aimed at reconciling different visions of the post-war future of the country, to create a political-administrative structure which would provide for a single polity, diminishing the centrifugal forces. Assignment of functions and sources of revenue between different tiers of government, as well as introduction of a transfer system, brought to the forefront numerous opportunities for various actors at local, cantonal, entity, and state level to pursue goals under the pretence of efficiency, redistributive role or macroeconomic stability. However, even though vertical and horizontal imbalances still exist in two BiH entities, throughout three phases they have been significantly improved. There is still space for further improvements especially with growing public debt issues which sets yet another challenge for the policymakers in BiH.

In this chapter, we evaluated the position of LGUs based upon available statistical data of 14 years. We can conclude that the position of LGUs has improved during this period. Even though the process of fiscal decentralisation in BiH has been very slow in comparison to other transition countries, we would say that it is becoming more successful. Our analysis has shown that the position of LGUs has strengthened in terms of higher democratisation at the level of LGUs—the responsibilities and authorities given to mayors and local councils in both entities, the LGUs’ negotiation power with the higher tiers of government improved and LGUs’ position in the revenue-sharing mechanism increased. The post-war reconciliation was mostly finished during early years in the first phase and that process was speeded up by the international aid for reconstruction and development. Challenges in BiH exist in the intergovernmental fiscal relations between higher tiers of government. As these fiscal relations continue to improve, we should expect direct benefits at the level of LGUs in both entities. In terms of local economic development, LGUs began to develop during the second phase when most capital investments took place and where positive effects of fiscal decentralisation were seen the most. However, GFC brought additional problems in terms of rising public debts and questions about the success of fiscal decentralisation which remains current challenge set for all tiers of government in BiH and not only local communities.