Journal Issue
Share
Article

Romania: Selected Issues

Author(s):
International Monetary Fund. European Dept.
Published Date:
March 2015
Share
  • ShareShare
Show Summary Details

More Fiscal Decentralization—The Prerequisites1

Core Questions and Findings

  • What is the current status of fiscal decentralization in Romania and how does it compare to peers? Decentralization in Romania started in 1991. The share of subnational government expenditure in general government expenditure doubled from about 12 percent in the mid-1990s to 25 percent in 2013 but is still below the European Union average.

  • What is the authorities’ plan to accelerate decentralization? A bold initiative was launched in 2013 to shift important responsibilities and assets of some ministries to local authorities. However, the Constitutional Court rejected the initiative. The government is now identifying smaller-scale pilot areas for decentralization, and intends to gradually expand the initiative.

  • What are the main weaknesses in the current system and potential risks from the plan going forward? The current system is characterized by a low share of revenue at the full discretion of local authorities, a large vertical imbalance, weak effectiveness of fiscal rules at the local level, a weak commitment control system, and a mismatch between central and local reserve funds. The plan going forward could worsen the vertical imbalance and jeopardize public assets, unless certain conditions are put in place.

  • How can Romania reap further benefits from decentralization and mitigate the associated risks? A stochastic frontier analysis indicates that there is room for Romania to improve the efficiency of its public expenditure. Such an improvement could allow larger and better growth-enhancing public investments without additional budgetary pressures. A cross-country analysis suggests that decentralization can improve efficiency but strong institutions and capacity are required. Local governments’ own taxes, such as property taxes, should be enhanced. Public financial management needs to be strengthened, including rolling out the new commitment control system to local public institutions and setting aside a budget buffer of 10 percent at the local level in line with the requirement for the central government. The effectiveness of fiscal rules at the local level should also be enhanced. Finally, strong accountability of local authorities to the local population is necessary. Absent such prerequisites, further decentralization can worsen public expenditure efficiency and macroeconomic performance.

A. Introduction

1. A shift to greater decentralization has occurred over the past two decades. The decentralization process started in 1991, after the end of the communist regime, with the adoption of the new constitution which established three layers of administrative structures. Various responsibilities were gradually devoluted from the central to local governments, including public utility services, local public transportation services, community services, and social protection in the 1990s; and health, education, agriculture, and local police thereafter. In the early 1990s, Romania was amongst the most centralized countries, as reflected by the share of local governments in general government’s revenue and expenditure. Tremendous efforts were deployed to double this share, bringing Romania in line with the average share of the Central, Eastern and South Eastern Europe (CESEE) countries. However, compared to many other European Union (EU) countries, Romania remains more centralized.

Subnational Government Expenditure, 1995–2013

2. A further decentralization initiative was put forward in 2013 but was rejected in the proposed form by the Constitutional Court. The draft law aimed at accelerating the implementation of the subsidiarity principle, i.e., bringing public services and decisions closer to the population and using resources more judiciously. It envisaged achieving this objective by shifting specific responsibilities and resources to the local authorities. The Constitutional Court argued that the draft law was not constitutional for a number of reasons: it lacked the required impact analyses and preparatory measures; cost standards for decentralized public services and related quality standards were missing; and the draft did not comply with the principle of local autonomy and with the constitutional property regime. Nevertheless, greater decentralization remains a key objective of the government; to address the Court’s concerns, the government has now opted for a gradual approach. Smaller-scale pilot areas are being identified, for which the requirements will be prepared (such as impact, cost and quality analyses). A specific law will be prepared for each pilot area and the initiative will be gradually expanded to other sectors to achieve the ultimate objectives.

3. This paper aims at providing policy options for the decentralization process. Section B presents the main merits and risks of decentralization drawing on existing literature. Section C describes the evolution and the current status of decentralization in Romania. Section D discusses the proposed system under the 2013 draft decentralization law. And Section E analyzes risks, mitigating measures, and options going forward.

B. Merits and Risks of Decentralization

4. There has been a trend towards decentralization internationally, but the degree differs across countries. Decentralization reflects country-specific characteristics and objectives. Historically, culturally and ethnically, decentralization has been used to protect the identity and independence of each group in the society or to create synergies among groups. Economically, decentralization is considered as a policy tool that can improve the provision of public services, and thereby, improve the overall macroeconomic performance. However, decentralization has been harmful in some cases where it was not well designed.

5. Decentralization can improve public service delivery through preference matching, stronger accountability, and competition. Policies devolved to lower-level governments are expected to better match the preferences of the local population and improve allocative efficiency (Hayek, 1945; Tiebout, 1956). When provided by the jurisdiction that has the control over the minimum geographic area, costs and benefits of public services are fully internalized (Oates, 1972). Geographical closeness of public institutions to the local population fosters accountability. Many studies found positive relationship between outcome of public services—particularly in education and health—and the degree of decentralization (e.g., as compiled by Ahmad et al., 2008).

6. In principle, decentralization of responsibilities and budget management could also improve fiscal and macroeconomic performances. The allocative and productive efficiency gains from decentralization can create budgetary savings and stronger fiscal performance as a result of lower costs of public service provision (Escolano et al., 2012; Neyapti, 2010; Baskaran, 2010). Decentralization encourages competition among local governments to attract tax base (businesses, property) through reasonable local tax burden, which leads to a search for innovative techniques to produce and provide public goods (cost-effectiveness), and reduces waste and corruption (Brennan and Buchanan, 1980). Local accountability pressures and better identification of the needs for public goods and services at the local level—including of infrastructure, health and education needs—can support economic growth (Fredriksen, 2013; Bénassy-Quéré et al., 2007; Kappeler and Valila, 2007; Arze del Granado et al., 2005).

7. However, if not carefully designed, decentralization can worsen the efficiency and quality of public services. If economies of scale are important in the production and delivery of a specific public good, decentralizing such responsibility to a small scale local government—in terms of population size—can increase production cost and reduce efficiency. In cases where the local authorities are not directly accountable to the local population, the preference matching and allocative efficiency might not materialize as the local authorities can allocate resources to less productive spending (Davoodi and Zou, 1998; Woller and Phillips, 1998; Zhang and Zou, 1998; Gonzalez Alegre, 2010; Grisorio and Prota, 2011). Accountability is weak, for instance, when local authorities are not elected by the local population but appointed by the central government. Moreover, the ability of the central government to redistribute resources, for example through equalization funds, and ensure minimum provision of public services to all population might be hindered if its share of revenue and expenditure is reduced (Ter-Minassian, 1997).

8. Decentralization could loosen fiscal discipline and jeopardize fiscal sustainability. Local policymakers may fail to internalize fully the cost of local spending when they can finance their marginal expenditure with central transfers or shared revenue that are funded by taxpayers in other jurisdictions. The marginal benefits of additional spending would exceed the marginal costs. Such a behavior, often referred to as the “common pool,” leads to overspending and deficit bias (Oates, 2006).2 The central government may not be able to enforce hard budget constraints on subnational governments that are consistent over time (Rodden et. al, 2003). Local governments may lack the capacity to efficiently manage the budget and provide public services. Finally, decentralization can increase financing costs as borrowing conditions are typically more favorable for central than local governments given that the former is perceived by the market as holding a privileged policy role.

C. The Current System in Romania

9. The share of “discretionary” own revenue is low and the share of earmarked funds has increased. Revenues at the local level comprise own taxes, shared taxes, earmarked and non-earmarked funds and transfers (see Annex I). The share of own taxes, for which local governments have full discretion on collection and use, declined and is now amongst the lowest in the European Union. However, it is to be noted that when shared taxes are included in own revenue, the share of the latter in total local government revenue expanded from about 20 percent of total revenue in the early 1990s to about 50 percent to date. Meanwhile, a system of earmarking of funds and transfers has gradually been introduced, and their share has exceeded 30 percent of total revenue of most local governments in recent years.

Local Government Revenue, 2012(Percent)
Revenue sourcesShare
Local taxes17.9
Allowances deducted from income tax30.6
VAT for financing decentralized expenditures23.6
Other transfers and other income17.5
VAT to balance local budgets4.9
Amounts allocated by the income tax4.3
VAT for roads0.9
Source: Romanian Authorities.

Sources of Subnational Government Revenue, 2013

(Percent of subnational government revenue)

1/Other includes property income, social contributions and other subsidies on production.

Sources: Adapted after European Commission 2012, Public Finances in EMU, Eurostat and IMF staff calculations.

10. The share of local governments’ expenditures has expanded but remains below that of peers and is largely non-discretionary. The share of local government expenditures in general government expenditures doubled in the past two decades. It has reached the average share in CESEE countries of about 25 percent in 2013, but is still below the EU average of about 33 percent. The education and health sectors represent the largest share of local governments expenditures (pre-school and primary education for counties and complemented by secondary and post high school for communes, cities and municipalities).3 Expenditures in those sectors lead to the predominance of items that are not at the full discretion of the local authorities (such as an important part of personnel and goods and services).

Local Government Expenditure, 2012(Percent)
Expenditure itemsShare
Personnel expenditures23.1
Goods and services23.4
Capital expenditures15.7
Interest subsidies and others37.5
Source: Romanian Authorities.

11. The vertical fiscal imbalance persists as the share of expenditures financed with own revenue remains low. Due to frequent changes in legislation, the magnitude of revenues at the full discretion of local authorities, i.e., own taxes, has oscillated significantly. Thus, an analysis of the vertical fiscal imbalance (VFI) based on such revenues requires particular caution. However, despite the large oscillations, the VFI has been on a worsening trend and Romania displayed one of the largest vertical imbalances across the European Union. The low VFI implies a heavy dependence on transfers from the central government. Furthermore, the frequent changes in legislation impede predictability and do not allow an efficient multi-year budgetary planning at the local level.

Share of Local Government Expenditure Covered by Own Taxes, 2013

(Percent)

Sources: Adapted after European Commission (2012), Public Finances in EMU, Eurostatand IMF staff calculations.

12. Local government budgets are bound by a balanced budget rule and a debt ceiling. Legislation requires that local governments’ budgets, excluding loans to finance investment and debt refinancing, must be balanced (“golden rule”). Moreover, local governments cannot contract or guarantee loans if their annual public debt service (principal payment, interest, commissions) including the loan they want to contract, is higher than 30 percent of their own revenue. From 2013, the budget of each government should be balanced, excluding investment projects financed by drawings on loans contracted before 2013. Loans contracted in 2013 were used to repay arrears.

13. Despite low local government debt, arrears have frequently accumulated. Local government debt is among the lowest in the European Union. However, many episodes of arrears accumulation occurred, particularly regarding EU-funded projects and in the health sector. The authorities deployed various ad-hoc measures to address this issue, including restricting local governments to contract new loans only to cover arrears payments and putting a cap on personnel hiring to allow budgetary room for arrears clearance. Hiring by local governments is based on population size.

D. The Proposal for Further Decentralization

14. The 2013 draft law aims at strengthening subsidiarity by identifying services to be shifted to local governments. The draft law indicates two main objectives: (i) to increase the quality and effectiveness of services provided to citizens by the transfer of powers from the central government to local authorities, and (ii) to strengthen the administrative capacity of administrative units to reduce disparities in economic development. While previous legislations were relatively broad, the 2013 draft law specifies a list of local directorates of key ministries that would be brought under the management of local authorities. This draft law also stipulates a timeline with a view to accelerate the decentralization process.4 Moreover, the draft law intends to improve EU funds absorption through a more fluid decision making process and stronger capacity at the local level.

15. In the proposal, revenue sharing would remain unchanged, and the new decentralized competences would be financed with additional transfers. No revision is envisaged for the methodology of sharing revenue between central government, counties, districts, cities, municipalities, and towns. The 2014 budget, which was adopted before the Constitutional Court took a decision on the decentralization legislation adopted by parliament, foresaw budgetary transfers from the ministries involved in the decentralization process to the local governments according to their respective additional competences. Those transfers would amount to about RON 674 million (or 0.1 percent of GDP), of which about half would be related to personnel expenditures and the other half to goods and services. The overall budget balance would not be impacted as the lower balance at the central level created by the transfers would be compensated by a similar higher balance at the local level.

Financing of Additional Decentralized Competences(Million lei)
Financing
Agriculture17
Culture2
Education182
Health310
Environment68
Sports94
Total674
in percent of GDP0.1
Source: Romanian Authorities.

16. More importantly, the draft law envisages the transfer of significant assets to local authorities. The draft law was prepared under the premise that full ownership of assets would enhance accountability in usage and maintenance. The assets to be transferred to local authorities include public transportation equipments and facilities, and sports and tourists facilities. The value of those assets has not been presented yet. Considering their magnitude, the implications of those asset transfers may be significant for the government balance sheet.

E. Policy Recommendations

Institutional environment

17. International experiences point to the need for a favorable institutional environment to reap the benefits from decentralization. A new empirical study by Sow and Razafimahefa (forthcoming) of 64 countries—including advanced, emerging and developing countries—over the period 1990–2012, analyzes the impact of fiscal decentralization on the efficiency of public service delivery (Annex III). They found that effective autonomy of local governments is required to allow preference matching and the allocative efficiency to operate. Strong accountability of local authorities vis-à-vis the local population, for instance through direct election, is necessary to allow improvement in productive efficiency. Corruption needs to be tackled to prevent any misuse of public resources. And capacity needs to be strengthened at the local level. Absent those conditions, fiscal decentralization can worsen public service delivery.

18. Romania’s efficiency of public expenditure lags behind that of peers. Sow and Razafimahefa (forthcoming) estimates the efficiency of public expenditure through a stochastic frontier technique (Battese and Coelli, 1988; Jayasuriya and Wodon, 2003; Grigoli and Kapsoli, 2013). This technique assumes that no economic agent (i.e., country) can exceed the ideal “frontier,” which refers to the optimum output—for instance infant mortality rate or school enrolment rate—produced with limited inputs (i.e., public expenditure). The deviation of the output in a specific country at a specific year from this frontier represents the individual measure of efficiency of that country. Efficient governments are those operating at, or very close to, the frontier as they try to reduce the infant mortality rate or improve school enrolment rate, given a limited amount of public expenditures. The model is specified as follows:

where Yit is public expenditure outcomes, namely the infant mortality rate (and alternatively) the secondary school enrolment rate, with subscripts i and t denoting respectively country and time dimensions; PEit-1 is public expenditure on health (and alternatively) education as percent of GDP; Zk,it is a set of control variables. The error term εit in equation (1) has two components as shown in equation (2); ωit represents an idiosyncratic disturbance, capturing measurement error or any other classical noise, and μit is a one-sided disturbance capturing the country-specific and time-varying efficiencies of public expenditures. This analysis indicates that the efficiency of public service delivery in Romania has improved over time but remains below that of countries in the region.

Efficiency of Public Expenditure

Source: Sow and Razafimahefa (forthcoming)

19. Provided that institutions and capacity are strengthened, Romania can improve the efficiency of public expenditure through further decentralization. Sow and Razafimahefa (forthcoming) analyzes the impact of decentralization on the time-varying and country-specific efficiency coefficients of public expenditure estimated above. The model is specified as follows:

where μit is the country-specific and time-varying efficiencies estimated from equations (1) and (2), fdit-1 measures fiscal decentralization (i.e., share of local government expenditure or revenue in general government expenditure or revenue, Iit-1 is a set of institutional and capacity variables, Wit is a set of other control variables and ψit is a stochastic error term. The set of institutional and capacity variables includes the level of corruption, the degree of autonomy of the regions, the strength of the democracy, bureaucracy, political stability, and checks and balances. The analysis finds that the fiscal decentralization variable alone worsens public expenditure efficiency. However, when interacted with stronger institutions and capacity, the impact becomes positive (Annex III). Thus, fiscal decentralization needs to be accompanied by measures to ameliorate institutions and capacity at the central and local levels. Romania has made progress on democracy and checks and balances, which is expected to support the positive impact of decentralization on public expenditure efficiency by fostering accountability. However, Romania lags behind peers on corruption and public financial management. Those weaknesses should be swiftly addressed before accelerating fiscal decentralization to prevent misuse or inefficient use of public resources which can lead to a deterioration of public service delivery.

Institutional Environment 1/

1/ Higher “democracy” indicator indicates stronger democracy; higher “corruption” indicates less corruption.

Vertical imbalance

20. The decentralization of resources and spending to local authorities should be carried out proportionately and simultaneously. In cases where decentralization of resources is accelerated without a matching decentralization of spending, the “excess” revenue can create additional spending. This can lead to a deterioration of the general government’s fiscal performance as the lower fiscal balance at the central level (due to lower revenue) is not compensated by higher balance at the local level. This case was witnessed in Colombia in the 1990s, Indonesia in 2000, and Nigeria in the recent oil-boom years. On the other hand, if responsibilities and spending are devolved without commensurate assignment of revenue, local authorities could be obliged to increase borrowing, accumulate arrears, or lessen the quality of public services. Transition economies in the 1990s experienced this phenomenon as they faced fiscal constraints at the central level and shifted responsibilities to local governments without proportionate resources.

21. Subnational governments’ own revenue, such as property taxes, should be enhanced to reduce the vertical imbalance. Romania’s low property tax rate can partly explain the low share of own revenue at the local level. Revenue collected from property tax is much below the EU average. Various analyses seem to suggest that there is room to increase the rate and raise revenue from property taxes, while protecting the poor against such an increase. Local authorities have full discretion over property taxes; they have legal latitude to change the parameters, as opposed to shared taxes. Assigning revenue responsibilities to local authorities would bolster collection efforts and strengthen fiscal discipline. This would enforce a hard budget constraint on local authorities as they would receive lower transfer from the central government, matching the assigned revenue responsibilities. Higher own revenue leads to more independence in spending allocation and allows the preference matching to operate. Insufficient own revenues would lead to over reliance on transfers from the central government and a tendency to spend without due consideration to the performance of revenue collection (i.e., a “soft budget constraint”). In addition to the property tax rate, own revenue can also be enhanced through the changes in the tax base (for instance, a shift from “ownership” to “use” principle or a more frequent reevaluation). Beyond the vertical imbalance, the horizontal imbalance should also be given sufficient attention to ensure that all subnational governments—the rich and the less rich—possess the minimum resources to provide all citizens standard public goods and services.

Public financial management

22. Strong public financial management, including commitment control and fiscal reporting, is essential to address risks from fiscal decentralization. Sound budget formulation, execution, accounting and control are critical. During the formulation and execution phases, local governments should set aside a fiscal buffer of 10 percent of total expenditure, in line with the requirement for the central government. This could help address potential mismatches between transfers from the central government and expenditures at the local level and prevent arrears accumulation. Local authorities should swiftly make use of the new commitment control and fiscal reporting systems. The new systems would contribute to improving management of local governments’ budgets based on real-time information at the various stages of the expenditure chain, bolstering transparency and efficiency. Training of users should be accelerated to address reluctance. Reporting on fiscal operations at the subnational level should be rigorous and timely. The investment prioritization initiative, recently launched at the central level, should be expanded to local projects. Financing through European funds should be sought before resorting to national financing to improve efficiency and reduce budgetary pressures. Finally, capacity at the local level should be strengthened. In case some local governments still lack the required capacity while others are deemed ready, decentralization of responsibilities and budget management can vary across subnational governments (asymmetric arrangements).

23. The value of assets that are potentially to be transferred to local authorities should be assessed and safeguard measures should be designed. As the 2013 decentralization draft law had envisaged assets transfers from the central to local governments, the value of those assets should be assessed. The value of the assets mentioned in the 2013 draft law seems significant, including transport and tourism facilities. The capacity needed for their management and the maintenance costs should also be analyzed. Absent such an assessment, the near- and medium-term quality of public services from those assets could deteriorate.

Economies of scale

24. Scale economies should be taken into account in the decentralization process as Romania ranks amongst the countries with a small population size per local government. Responsibilities to be shifted to local governments should be thoroughly selected. Efficiency could worsen and production costs could increase if scale economy is important in the production and delivery of the public goods/services. The size of population at both the local level (city, town or municipality) and intermediate level (county) is smaller than the EU average.5 Fiscal decentralization should be driven by efficiency and capacity considerations, and the role at each level of government should be clearly defined. Some overlapping of task assignments has been noted in the current decentralization legal framework. Based on those considerations, the transfer of the management and ownership to local authorities of some large transport and tourism facilities as envisaged under the 2013 decentralization initiative should be cautiously analyzed.

Average Population Size at the Local Level

(Thousands)

Sources: Council of European Municipalities and Regions, World Bank and IMF staff calulcations.

Average Population Size at the Intermediate Level

(Millions)

Sources: Council of European Municipalities and Regions, World Bank and IMF staff calulcations.

Fiscal rules

25. The effectiveness of fiscal rules for local governments could be strengthened to mitigate risks from decentralization. Fiscal rules for local governments in Romania are broadly similar to that in peer countries.6 However, the enforcement framework is significantly weaker as Romania ranks poorly on the European Commission’s Fiscal Rule Strength Index at the local level. This index informs on (i) the legal base of the rule (such as constitution or law), (ii) the room for revising the objectives, (iii) the institutions in charge of enforcement, monitoring and alerting (such as independent authority or the parliament), (iv) enforcement mechanisms (such as sanctions and automatic corrections), and (v) media visibility (triggering public debate). Romania lacks a mechanism that would trigger an alert in case of deviation from the rules; the actions of the recently created Fiscal Council have so far been limited to the general government but have not sufficiently covered the local governments’ fiscal rules yet. Also, effective corrective actions and sanctions in case of non-compliance are insufficient, despite some controls by the Court of Accounts and some penalty provisions. For instance, in the province of British Columbia in Canada, ministerial salaries can be withheld; and in Brazil, officials may be subject to fines.

F. Conclusions

26. There seems to be room for Romania to improve public expenditure efficiency and growth-enhancing investments from further decentralization. The efficiency of public expenditure seems markedly below countries in the region and the EU average. A cross-country analysis suggests that decentralization can improve efficiency but strong institutions and capacity are required. Romania has accelerated its decentralization in the last two decades as the shares of subnational government revenue and expenditure in general government revenue and expenditure doubled between 1995 and 2013; however, it remains more centralized than the EU average. Improving public expenditure efficiency can allow larger and better public investments without additional budgetary pressures.

Fiscal Rule Strength Index

Source: European Commission.

27. However, a number of prerequisites need to be in place to reap those benefits. Local governments’ own taxes should be expanded to reduce the large vertical imbalance; an enhancement of the base and rate of property taxes could be considered. Sanctions against non-compliance to local fiscal rules could be designed and enforced; media visibility could be strengthened and alert mechanism could be put in place. The new commitment control system should be swiftly used by all public institutions at the central and local levels. Local governments could be encouraged to set aside sufficient fiscal buffers in line with that at the central government level. The value of the public assets that could be transferred to the local authorities should be assessed and adequate safeguard measures put in place. Finally, corruption should be tackled to prevent efficiency leaks from further decentralization. Absent those prerequisites, further decentralization can worsen public expenditure efficiency and macroeconomic performance. Some lessons learned from experiences of other countries are shown in Annex IV beyond the issues covered above. Such experiences include excessive revenue decentralization, overlapping responsibilities between layers of governments, tax competition between local governments, and rigid earmarking of transfers.

Annex I. Subnational Government Revenue Sharing

Local governments’ revenue consists of own taxes, shared taxes, and transfers from the central government, as follows:

  • Income tax for the case of Bucharest Municipality: 20 percent to the budgets of Bucharest districts, 44.5 percent to the budget of Bucharest Municipality, 7 percent for balancing the district budgets and the budget of Bucharest Municipality, and 28.5 percent for the central government;

  • Income tax in other cases: 41.75 percent to the local budgets of communes, cities and municipalities where the income tax payers carry out their activities, 11.25 percent to the local budget of the county, 18.5 percent to balance the budgets of communes, cities and municipalities, as well as the budget of the county (equalization fund), 28.5 percent for the central government;

  • Profit tax from companies under the authority of county councils;

  • Equalization transfers: transfers from certain revenues in the state budget shared to counties as follows: (i) 70 percent according to the financial capacity of the county (based on the distance of income tax per capita collected from the average collection in all counties, and the number of inhabitants), (ii) 30 percent according to the county area;1

  • Property taxes (taxes on building and taxes on land); charges from transportation, permits and licenses; income from services; fines and penalties; hotel duty and tax from entertainment; sale of assets; donation and sponsorship; financial operations;

  • Subsidies from the state budget; and

  • European funds.

Annex II. Subnational Government Expenditure Composition

Share of Subnational Government Expenditure by Sector, 2012

(Percent of general government expenditure by sector)

Source: Eurostat.

Share of Subnational Government Expenditure by Economic Function, 2013

(Percent of general government expenditure by economic function)

Source: Eurostat.

Annex III. Decentralization, Efficiency and Institutions
Table AIII.1:Fiscal Decentralization and Efficiency of Public Service Delivery
Dependent variable: estimated efficiencies
HealthEducation
Variables(1)(2)(3)(4)(5)(6)(7)(8)
FD(t-1)-0.523-0.809-1.307***-0.727***-1.079*-0.171-0.764*-0.696
(-1.540)(-1.137)(-2.703)(-3.159)(-1.780)(-0.217)(-1.889)(-1.275)
FD × Corruption(t-1)-0,488***-0,608***
(-3.291)(-2.738)
FD × Parliamentary(t-1)4.373**1.160
(2.206)(0.836)
FD × Regime(t-1)0.033***0.0125
(2.967)(1.477)
FD × Autonomy(t-1)2.057***1.952***
(5.457)(2.931)
Real GDP pc(t-1)-0.040-0.122-0.117***0.013-0.130**-0.044-0.0717**-0.020
(-1535)(-1.598)(-2.803)(1.154)(-2.371)(-0.920)(-2.257)(-1.120)
Note: (*), (**) and (***) denote statistical significance level of 10, 5 and 1 percent respectively. Robust t-statistics in parentheses.Source: Sow and Razafimahefa (forthcoming)
Table AIII.2:Fiscal Decentralization and Public Expenditure Efficiency: Alternative Political and Institutional Variables
Dependent variable: estimated efficiencies
HealthEducation
Variables(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)
FD(t-1)-0.486*0.953***-1.408-0.480***-0.022-0.0070.786**-0.8110.565-0.014
(-1.838)(3.609)(-1.513)(-2.741)(-0.061)(-0.003)(2.143)(-1.310)(1.519)(-0.008)
FD × Assembly elec.(t-1)3.672***5.499
(3.093)(0.525)
FD × Presidential(t-1)-1.737***-1.410***
(-4.999)(-2.583)
FD × Bureaucracy(t-1)0.3790.16
(0.953)(0.644)
FD × Political stab.(t-1)0.1020.459
(0.781)(1.394)
FD × Checks and balances(t-1)0.141-1.032
(0.924)(-1.216)
Real GDP pc(t-1)0.054***0.002-0.0080.0060.0090.023-0.034**-0.032-0.0220.080
(5.817)(0.169)(-0.282)(0.412)(0.540)(0.786)(-2.319)(-1.541)(-1.065)(0.910)
Note: (*), (**) and (***) denote statistical significance level of 10, 5 and 1 percent respectively. Robust t-statistics in parentheses. Source: Sow and Razafimahefa (forthcoming)
Annex IV. Lessons Learned

Excessive revenue decentralization: China decentralized the revenue sharing arrangement in 1980s, which led to a sharp drop in the central government’s share in general government revenue from about 40 percent in 1985 to 28.3 percent in 1993 (IMF, 2009b). By the early 1990s, this share became critically inadequate as the central government could no longer conduct sufficient redistributive policies. China’s impressive growth performance has not benefited all subnational governments equally, as income disparities across provinces have widened. Furthermore, the provision of public services is skewed in favor of richer provinces. These developments called for larger redistribution. In 1994, a set of reforms was introduced to change the revenue-sharing arrangement, create a central tax administration and increase the role of transfers. By 2007, the share of the central in general government revenue reached 53.5 percent.

Overlapping responsibilities: In Bolivia, significant overlaps in spending responsibilities created inefficiencies and reduced accountability (IMF, 2009b). Health and education were characterized by extensive concurrency between central government, regions, and municipalities. The central government set the norms and the curricula, and pays for salaries of teachers and medical personnel. Municipalities were responsible for construction and maintenance of school and health premises and for educational and medical equipment and supplies. Regional governments were responsible for the implementation of norms and standards set by the center. Selection and hiring of teachers and medical personnel was done at the regional level, but the certification of hours worked is done at the municipal level. Lack of coordination led to spending inefficiencies; one of the consequences of this setup was that, while new construction generated new requests for personnel to run the premises and to provide the services, separation of responsibilities originated frequent discrepancies between building of new premises and their staffing. More generally, investment spending decisions were not coordinated across levels of government.

Tax competition: In Brazil, the Tax Incentive War (“Guerra Fiscal”) among local governments is a manifestation of a “race to the bottom” (Castilho and Silveira, 2009; Ayres dos Santos, 2013). States used the ICMS, the most important tax in Brazil which is somewhat similar to VAT, to design schemes that would provide tax incentives to attract businesses (trade and investment). Although, tax competition is unconstitutional, states found various ways to circumvent the legal framework, including deduction and exemption schemes. The tax competition bitterly penalized small states who did not possess sufficient tax lever to provide tax incentives and were not endowed with competitive and large production base. This led to large disparity between states in provision of public goods and services and in level of development. A degree of coordination of subnational taxes, especially as regards the definition of the tax base, is important to avoid predatory tax competition. Tax competition can also lead to large vertical imbalance. The decentralization framework should give incentives and possibilities to local authorities to raise the necessary revenue.

Rigid earmarking of transfers: In Kosovo, formulas for earmarked transfers in health and education were not adequate to cover municipalities’ spending mandates in these areas, while unconditional transfers are not providing any spending autonomy to municipalities, given their limited size. In Mexico and Indonesia, special-purpose earmarked transfers translated into rigidity in subnational spending and the effective allocation to the specified purpose was rather difficult to control. Earmarked transfers should be carefully designed (IMF, 2009b).

References

Prepared by Ivohasina F. Razafimahefa.

Moreover, the conduct of countercyclical policies could be weakened if a large share of taxes and spending is shifted to subnational governments as the central government would not have sufficient policy lever and policy priorities differ across levels of government.

Annex II presents a comparison by sector and economic function in Romania to the averages for the EU.

From the mid-2000s, some discussions were launched to create a new administrative layer—the “regions” —between the central government and the counties. The regionalization initiative has not been implemented as it would require an amendment of the Constitution for which political support has been insufficient. This debate slowed down the decentralization process.

As some European countries have higher number of administrative layers than Romania, the intermediate level in this analysis corresponds to the administrative layer immediately following the lowest layer.

See European Commission’s fiscal rules database.

From the income tax equalization fund and the equalization transfers: 27 percent to county’s budget, and 73 percent to the budgets of communes, cities and municipalities distributed as follows: 80 percent decided by the head of the county public finance general directorate, based on population, land area and financial capacity, and 20 percent decided by the county council to clear arrears.

Other Resources Citing This Publication