Abstract
This paper describes the institutional changes that have induced a decline in the vertical fiscal imbalance (VFI) - defined as the share of sub-national own spending not financed through own revenues - in four European countries: Belgium, Italy, Norway, and Spain. The decline in VFI was achieved through progressive devolution of revenues to sub-national governments in Belgium, Italy, and Spain, while re-centralization of health sector expenditures was the cause of the decline in the VFI in Norway.
I. Introduction
Vertical fiscal imbalances (VFIs) have often been identified as an obstacle to sub-national accountability and good fiscal performance. VFIs occur when spending decentralization outpaces revenue decentralization. In these cases, sub-national governments must rely on central government transfers and borrowing to finance expenditures. In breaking the link between sub-national service provision and its financial consequences for the electorate, VFIs are commonly viewed as being responsible for sub-national fiscal profligacy and soft budget constraints. Recent econometric evidence (Eyraud and Lusinyan, 2011) confirms that increases in VFIs are associated with a weakening of general government fiscal balances.
VFIs come into greater scrutiny during difficult times. Attention to VFIs increases when deficits become chronic, necessitating frequent bailouts. Sub-national bailouts can put intergovernmental fiscal relations under severe strain, providing ground for re-centralization as a means to eliminate the VFI. Changes in relative fiscal positions—for example, when central government revenues thrive while those at the local level languish—also tend to put VFIs into the limelight. This is especially critical in Europe, where the ongoing financial crisis has exacerbated fiscal pressures at all levels of government. Understanding how changes in VFIs affect performance is thus of critical importance for the design of fiscal adjustment strategies, especially those that contemplate changes in intergovernmental relations as part of their fiscal reforms.
The case studies identify periods of changes in intergovernmental relations that have induced variations in the VFI. The approach complements the cross-country econometric evidence of Eyraud and Lusinyan (2011). In each of the countries considered, we study decentralization policies that have brought about declining VFIs, as well as the economic environment and external factors that may have contributed to changes in fiscal balances. The countries Belgium, Italy, Norway, and Spain were selected because they have experienced a significant decline in VFI over the past two decades.
The remainder of this paper is structured as follows: Section II describes the measure of VFI; Section III presents four case studies on Belgium, Italy, Norway, and Spain; and Section IV draws some lessons from the case studies.
II. The Vertical Fiscal Imbalances
There is no consensus in the theoretical and empirical literature on the definition of the VFI. The most commonly used indicator in the empirical literature links VFI to the concept of transfer dependency. However, transfers are sometimes reported as a share of sub-national spending (Jin and Zou, 2001), sub-national total revenues (Rodden, 2002; Baskaran, 2010), or central government revenues (Bahl and Wallace, 2007). Often authors measure VFI simply as the difference between own revenues and own spending (Bird and Tarasov, 2004). More complex indicators that attempt to introduce a measure of actual tax and expenditure autonomy are qualitative and therefore subjective in nature. Country case studies focus mostly on the asymmetry between revenue and expenditure decentralization and describe the economic and political circumstances during which the decline in the VFI occurred.
The measure of VFI in this study is an expansion of the most commonly used indicator. The VFI is borrowed from Eyraud and Lusinyan (2011) and defined as the share of sub-national own spending not financed through own revenues. By definition, the counterparts of VFI are sub-national borrowing and transfers received from other units of general government—both expressed as shares of sub-national own spending. The indicator measures the mismatch between spending and revenue decentralizations that widens when countries devolve more spending than revenue.
Following Eyraud and Lusinyan (2011), we define vertical fiscal imbalances as follows:
In this context, sub-national government (SNG) spending can be financed from either SNG own revenue, transfers received by the SNG, or borrowing:
The VFI can be altered through direct decentralization decisions, changes in sub-national fiscal policy, and changes in intergovernmental relations that strengthen sub-national fiscal discipline. The VFI widens when countries devolve more spending than revenue to sub-national governments, but also varies in response to events not related to decentralization decisions, for example, when sub-national spending increases and/or when sub-national revenue effort weakens, both leading to higher transfers and/or borrowing. Changes in intergovernmental relations that lead to greater discipline at the sub-national level can also alter VFI. For instance, a newly awarded power to change tax levels may induce discipline (making a no-bailout policy credible). While such changes would not affect sub-national revenue shares, they would be reflected in the VFI to the extent that they result in lower transfers and borrowing.
This measure of the VFI, like other indicators based on actual revenue and spending figures, is a less-than-perfect measure of sub-national government autonomy. The main shortcoming is that the VFI does not accurately capture the expenditure -and revenue-raising “powers” of sub-national governments (Rodden, 2002; Blöchliger and King, 2006; Bach and Wallau, 2009). Sub-national governments’ “discretion” over revenues may be overstated when tax sharing is recorded as own taxes, or when discretion over tax rates and bases is limited, or when the powers granted to sub-nationals are, for a variety of reasons, not used. On the spending side, a large share of sub-national spending may be regulated or earmarked in some arrangements where sub-national governments de facto act as de-concentrated agents of a central government that dictates levels and standards of service. This is frequently the case in health care, where central and sub-national governments also retain overlapping responsibilities.
Relating measures of the VFI to general government fiscal performance is challenging and beyond the scope of this paper. Sub-national deficits are part of the general government balance. Thus, part of the changes in the VFI indicator is attributable to changes in the sub-national deficit/surplus. At the general government level, fiscal balances are correlated with the VFI, but the causality may go both ways: widening the VFI may be a cause as well as a consequence of general government fiscal deficits. These challenges can best be addressed within an econometric framework such as the one adopted by Eyraud and Lusinyan (2011). Here, we make use of a case study approach to understand how VFIs have declined in practice through a more in-depth consideration of the economic environment and institutional circumstances of countries experiencing changes in the VFI.
III. Case Studies
A. Belgium
The Belgian federation includes three levels of government: the federal government, the regions and communities, and the provinces and communes. There are three geographical regions (Flanders, Wallonia, and Brussels-Capital) and three language communities (Flemish, French, and German). Regions and communities have their own governments and are not subordinated to the federal government in matters of expenditure policy. The federal government is responsible for common functions such as the judiciary, foreign policy, defense, and social security. This system was the result of constitutional changes that occurred between 1970 and 1993.
The large VFI in Belgium, between 60 and 70 percent over the last two decades, stems from strong revenue centralization at the federal level. Transfers comprise a limited number of grants. Communities do not have own taxes; they are financed mainly through general-purpose grants. Regions enjoy taxing power and receive federal grants. Yet nearly 80 percent of their funding consists of transfers, and only 20 percent from own taxes.
Decentralization and the VFI
Expenditure decentralization in the 1990s increased VFI, while revenue decentralization of the 2000s decreased it. Three legislative acts—the 1980 Act of Regionalization, the Special Act of August 1988, and the Lambermont 2001 agreement—have been responsible for the most pronounced changes in the VFI.
The 1980 Act of Regionalization introduced a limited devolution of revenue-raising powers. Of the three legislative changes, this is the one that had the lowest impact on sub-national own revenue shares and thus on the VFI.
This was followed by large-scale expenditure decentralization in 1988–89, during which the communities took charge of education spending while regions were entrusted with regional development and investment spending. On the revenue side, the reform introduced a new taxsharing system under which taxes collected by the federal government were transferred to sub-national governments by means of a set of “repartition keys” (IMF, 2003).2 While own sub-national revenues picked up in this period the VFI increased as the transfer of expenditure responsibilities outpaced revenue devolution (see Figure 1).
In 2001, a strong push for greater revenue devolution took place. A number of new taxes were transferred to the regions and the shared PIT margin was increased; VAT transfers to communities increased, and the revenue-sharing mechanism was revised to dampen the volatility of transfers to regions and communities.3 The 2001 reform increased the share of sub-national own taxes, reducing the VFI.
Belgium: Decentralization, VFI, and Fiscal Balances, 1985–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance: share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Belgium: Decentralization, VFI, and Fiscal Balances, 1985–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance: share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Belgium: Decentralization, VFI, and Fiscal Balances, 1985–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance: share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.With the narrowing of VFI, the flow of federal transfers and tax sharing declined while equalization grants substituted for block transfers. In line with the increase in tax sharing in 1989, block grants declined and inter-regional equalization was introduced in 1990 to compensate regions with lower per capita tax revenues. The federal government was allowed to compensate for the devolution of some taxes by cutting the amount of shared taxes with the regions. Since the amount of transfers to the regions is based on the “territoriality principle,” in 2002 an equalization transfer was allocated to the regions that have per capita personal income tax revenues below the national average (Verdonck and others, 2009).
Revenue devolution had a positive yet small effect on the VFI. With the doubling of sub-national own expenditures in 1989, the VFI widened by 15 percent (Figure 1), narrowing down only slightly over the next two decades with the gradual devolution of revenues. After the 2001 devolution of taxing power, regional taxes have increased by a couple of percentage points of GDP, lowering the VFI correspondingly by a few percentage points. However, by 2008, the VFI was still very high at 59 percent.
Revenue devolution in Belgium has been associated with an improvement in fiscal outcomes. Both the federal government and the sub-national entities’ fiscal positions have improved in the 1990s and early 2000s, turning the large deficits of the 1980s into a surplus in 2000. Public debt declined from 130 percent of GDP in 1993 to 80 percent of GDP in 2007.
The federal government contributed the lion’s share of the consolidation. Although regional (direct as well as indirect) tax revenue growth was strong and generally more dynamic, particularly during the 1990s and with the 2001 devolution, the federal government achieved a substantially bigger primary surplus because of across-the-board expenditure restraint. The debt decline accomplished in the 1990s consolidated the federal and general government’s overall fiscal position in the 2000s, thanks to positive spillovers from declining debt service outlays and robust economic growth.
B. Italy
There are three levels of sub-national government in Italy—the regions, the provinces, and the municipalities—with no hierarchical links. Similar to Spain, decentralization is asymmetric. Italian regions are divided in two groups, “ordinary” and “special” statute. The five special regions, which were introduced by the 1948 Constitution in response to a threat of secession, were granted extensive spending autonomy and central funding. The 15 ordinary regions were set up only in 1970, when the first regional elections took place. A second level of asymmetry was introduced by the 2001 Constitution, which allowed temporarily differing degrees of autonomy in different regions, but up till now it has not been used. Italian regions differ to a large extent in terms of size, population, and economic development.
Decentralization and the VFI
While decentralization has been a long ongoing process in Italy, a few historical periods with distinct changes in intergovernmental relations that may have affected the VFI can be identified. Some major institutional changes occurred as a result of the economic crisis. They aimed to strengthen local and national finances, harden the budget constraints on lower levels of government, and increase local government accountability. These periods, which induced a profound shift in intergovernmental relations, include the economic crisis of the 1970s, the political crisis of the 1990s, and Italy’s preparations to join the EMU in 1998.
1970s and 1980s: Most functions related to health care were assigned to the ordinary statute regions, as well as agriculture, industry and local transport. Local taxing powers were heavily centralized in the early 1970s by a massive tax reform. Forgone sub-national revenue was replaced with a system of earmarked transfers and tax sharing from the national government. In this period, sub-national government spending systematically exceeded own revenues, and the deficit was financed ex post by the central government (Bordignon, 2000). Sub-national own revenues were only 2 percent of GDP in 1980, while the VFI reached 80 percent.
1990s: A broader liberalization took place with reforms involving further expenditure decentralization. On the revenue side, new local taxes were assigned to regions and municipalities, with some latitude allowed in setting tax rates for local needs.4 There was an increase in block grants, and some regional sharing of central government taxes and own taxes. Reforms also occurred on the electoral front in an effort to improve the accountability of local politicians at the municipal level (1993) as well as at the regional level (1999).5 In particular, the electoral system was changed from being a proportional (parliamentary) system to a majority (presidential) one. Revenue from taxes of sub-national governments increased sharply, from 3½ percent in 1990 to 8½ percent of GDP in 2000,6 while the VFI shrank from 75 percent to 40 percent in just over one decade (Figure 2).
2001 onward: The Constitution was altered in the direction of greater decentralization, while also allowing for “asymmetric federalism.” Executive power was transferred to municipalities and provinces, allowing them to introduce new taxes while ruling out transfers by the central government as a method of ex post deficit financing. Moreover, an interregional redistribution fund was introduced to equalize revenue generation capacity across regions. Sub-national governments were allowed to borrow to finance investment expenditures. Still, areas of overlapping competencies were maintained, giving rise to conflicts across government levels. Regions’ new tax legislative powers were, however, not realized in the following years because of political obstacles (Ambrosanio and others, 2008). Hence, the VFI stabilized around its 2000 level.
Italy: Decentralization, VFI, and Fiscal Balances, 1980–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Italy: Decentralization, VFI, and Fiscal Balances, 1980–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Italy: Decentralization, VFI, and Fiscal Balances, 1980–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.The highest degree of decentralization was achieved in the health sector. The social and health sectors represent the principal spending functions of the regions and account for 80 percent of their total spending.7 Health care spending is managed at the regional level through the Local Health Authorities (LHAs). The specific aim of the LHAs is to guarantee, across the national territory, essential levels of treatment determined by the central government in cooperation with the regions (Buglione and Maré, 2010).
The decentralization reforms of the 1990s that shrank the VFI significantly were associated with improvement in the fiscal balance. Strong fiscal consolidation occurred first on the revenue, then on the expenditure side, with the general government primary balance swinging into a surplus exceeding 4 percent of GDP between 1992 and 1995. The adjustment was heavier on the revenue side in 1992 and 1997 when a tax on real estate and a temporary levy on personal income (the Eurotax) were introduced respectively. On the expenditure side containment of the wage bill through limits on hiring at central government level, the reform of pensions, and cost containment in health care all contributed to lowering deficits, although the latter was short lived (Balassone and others, 2008).8
C. Norway
Norway has four levels of government: central, regional, county, and municipal. Sub-national governments are under strict control of the center; indeed, the system is often referred to as administrative federalism, wherein local and county governments are de facto central government agents. Central government grants constitute a substantial part of sub-national revenue and local taxes are highly regulated by the center (Borge, 2006). Welfare services—education and health—were delegated to sub-national governments rather than being decentralized until 2002 (Rattsø, 2003), in line with trends throughout continental Europe.9 Over the last decade the central government role has strengthened and in a number of northern European countries, including Norway, the health sector was recentralized.
Decentralization and the VFI
Changes in the health sector have played an important role in changes in VFI. The health sector in Norway accounted for almost 10 percent of GDP in 2009. Before 2002, approximately 32 percent of sub-national spending was dedicated to health care. Norway’s health spending is one of the highest among OECD countries in per capita terms, right after the US and Switzerland. Public spending on health was broadly stable over the 1990s at about 7 percent of GDP. In the 2000s it has increased by almost one percent of GDP.10
Responsibility for health care planning and delivery has shifted in opposite directions over the last three decades in Norway.
Starting in 1969, service delivery was formally decentralized, though tied to central planning and financing.11
The first wave of centralization of service provision took place in 1974 with the establishment of five geographical regions aimed at exploiting economies of scale. Conflicts pervaded both the relationship between the counties and the central government, and between hospital administrative and professional management. These conflicts were believed to have undermined efficiency, equity, and cost control, leading to a blame game among counties (Magnussen and others, 2006).
As it became evident that counties were not cooperating voluntarily, regional coordination of health plans was made mandatory in 1999. Subsequently, in 2002, health services were fully recentralized as part of a larger hospital reform effort. The central government took over the responsibility of public hospitals and specialist care from the counties. The health regions were transformed into regional health enterprises (RHEs) in charge of purchasing health care from public or private providers, and organized as enterprises or trusts (Minas, 2010).12
In an attempt to control deficits while providing equitable services, health care financing modalities were also altered a number of times. Centralized financing was deemed necessary from the beginning because of inadequate tax bases for local governments, with property taxes and user charges accounting for a small share of own expenditures. From 1980, the health system was financed via central government retrospective reimbursements, later replaced with block grants. In 1997, large-scale activity-based financing was introduced, and the block grant was made proportional to the number of hospital treatments. In light of very different income levels across urban/rural regions, large equalization transfers were also provided (Borge, 2006).13
The evolution of the VFI was affected by the devolution and subsequent recentralization of health care, although other factors have also been important.
The increase in expenditure decentralization between 1996 and 1998 (first panel) was associated with a decline in the VFI (second panel). This was due to a sharp increase in the share of sub-national revenues in total general government revenues (first panel), which was the consequence of declining central government revenues in percent of GDP (fourth panel).
Re-centralization of health care in 2002 shrank the VFI in the following two years. However, the decline in sub-national own spending on health was more than offset by higher transfers and the VFI subsequently increased. This occurred as recentralization and fiscal expansion took place concomitantly.14
D. Spain
Spain is a classic case of asymmetric decentralization that happened in an incremental way. Spain has four tiers of government—the State, the 17 “regions” called Autonomous Communities (ACs), the provinces, and the municipalities.15 The ACs were established by the democratic Constitution in 1978 which introduced two levels of asymmetry. First, recognizing strong regional identity differences, the Constitution granted the historical regions of Basque Country and Navarre a separate “foral” regime AC status. These two regions have had a large degree of autonomy over tax collection that they have retained, almost unaltered, throughout most of their history.16 Second, the Constitution distinguished between regions that are placed on a “fast track” and those that would be on a “slow track” of decentralization (Molero, 2001). The targeted level of expenditure autonomy and broadly similar responsibilities across ACs were achieved gradually by 2003, when a complete transfer of health and education services to “common” regime ACs took place.
Decentralization and the VFI
The decline in the VFI was achieved gradually through two main reforms that took place in 1997 and 2002. Spain is one of the few OECD countries that experienced a large increase in own tax revenues as a percent of total sub-national government revenues between 1995 and 2007 (Charbit, 2009). Until 1996, the ACs relied substantially on government grants or tax-sharing arrangements with own revenues accounting for up to 16 percent of general government receipts (“decentralization” in Figure 3), as their tax competencies were confined to collection and administration. When further tax autonomy was granted in 1997, the own revenue share in general government revenues increased to 19 percent. The decline in the VFI was more pronounced, shifting from 57 percent of GDP to in 1996, to 49 percent in 1999, in line with increased devolution and declining sub-national deficits.17 Sub-national revenue shares increased more drastically following the 2002 devolution, rising to 13.2 percent of GDP in 2006, with the VFI declining to 30 percent of GDP.18 At end-2009, a new financing system approved for the common regime ACs transferred further resources increasing their revenue autonomy, and granted more discretionary powers on the composition and volume of revenues (IMF, 2010).19
Norway: Decentralization, VFI, and Fiscal Balances, 1996–2009
(Percent of non-oil GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Norway: Decentralization, VFI, and Fiscal Balances, 1996–2009
(Percent of non-oil GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Norway: Decentralization, VFI, and Fiscal Balances, 1996–2009
(Percent of non-oil GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.The transfer system was also modified to match the new spending and revenue responsibilities and address regional VFIs. The VFI was financed in the beginning of the 1980s through general-purpose grants. With a gradual increase in the role of tax sharing, transfers were reformed into: (i) a general grant—in the form of a share of central government general revenues—for financing common services and education; (ii) earmarked grants, e.g., for health and social services; and (iii) the Inter-Territorial Compensation Fund (ICF), designed to compensate for inter-regional differences and fund local capital expenditure (Mora, 2009).20 In 2002, as health services were transferred to ACs, regional health care financing was split from the rest of the transfers, by consolidating health care funding within the regional general revenue allocation system (López and others, 2004). The system of earmarked grants was replaced by a rule targeting per capita equity and the volume of transfers increased to match the ACs’ actual spending.
Declining VFI coincided with improved fiscal performance, though against the backdrop of a favorable economic context. The general government’s overall fiscal position strengthened between 1997 and 2007. However, a continuous economic expansion beyond the country’s long-term potential—with buoyant revenues from property boom and lower interest rates on debt following EMU accession—supported general government fiscal consolidation (Miaja, 2005). Meanwhile the transitory positive revenue surprises allowing real increases in expenditure, masked structural deficits.
Spain: Decentralization, VFI, and Fiscal Balances, 1995–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Spain: Decentralization, VFI, and Fiscal Balances, 1995–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.Spain: Decentralization, VFI, and Fiscal Balances, 1995–2009
(Percent of GDP, unless otherwise stated)
Citation: IMF Working Papers 2012, 091; 10.5089/9781475502664.001.A001
Sources: OECD database; and IMF staff estimates.Notes:1 Vertical Fiscal Imbalance (VFI): share of subnational own spending not financed by subnational own revenue. Transfer Dependency (TD): net transfers received by subnational governments as a share of subnational own spending. Balance: net borrowing as a share of subnational own spending.IV. Lessons from Case Studies
As the very long decentralization histories in Italy, Spain, and some other countries like Australia show; decentralization tends to be slow, incremental, and can experience setbacks. Declining VFI is difficult to achieve, as revenue devolution is sometimes resisted in decentralized spending settings. Some of the obstacles to devolution of revenue to sub-national governments include:
Loss of “cheap” resources. Less populous and poorer local governments are sometimes halfhearted with respect to devolution of revenue responsibilities, as they fear a decline in equalization grants and associate it with higher disparities (Blochliger and Petzold, 2009). Indeed, some Spanish ACs—Andalucia, Castilla-La Mancha, and Extremadura—initially opted out of the 1997–2001 financing agreement that granted some authority over rates of ceded taxes and an additional 15 percent of the income tax (Garcia-Mila and McGuire, 2002).
Inadequate tax base. Some taxes with broader bases, such as the PIT and CIT, are less adequate for sub-national governments as are those with re-distributive properties and mobile. This limits increasing levels of tax autonomy that can create inefficiencies, ultimately eroding overall revenues.
Weak administration. Administrative capacities for carrying out increasing revenue responsibilities may not exist. Administration and poor data quality and accounting practices have been considered obstacles to greater decline in the VFI in Italy during the 1990s and beyond (Balassone, 2008).
Tax competition. The international tendency towards lowering the tax burden, and the need to support the economy during downturn, coupled with the threat of fiscal mobility, do not bode well with the need to exercise increasing fiscal autonomy by raising sub-national taxes.
One of the possible solutions for addressing unequal regional preparedness for revenue devolution, both developmental and administrative, is the so-called “multispeed federalism.” The latter is usually associated with linguistic and ethnic differences, separatist tendencies, and generally, history. Italy and Spain both have five special statute regions that had been given greater autonomy early in the decentralization process for historical and political reasons.
Spain’s 1978 Constitution has introduced two decentralization “tracks,” allowing some regions to achieve full expenditure autonomy at a faster pace. Expenditure decentralization was matched by increasing (although incomplete) revenue devolution, widening the VFI.
The sequencing of decentralization may also vary across regions, as was the case in Italy. Special statute region experienced a negative VFI early on as revenue devolution was larger than the transfer of expenditure competencies.21 While Italy has opened up the door for a two-speed process for revenue devolution to ordinary statute regions in 2001, such an option has not been implemented, as of yet. Thus, different VFI across sub-national governments may co-exist over a certain period, affecting the aggregate VFI.
Some disadvantages to the multispeed federalism are the potential sacrifice of service uniformity, a possible loss of momentum for reform in regions that are lagging in the process, and increasing complexity in the transfer system in the intermediate period. This complexity represents a tradeoff with the greater accountability that can be achieved by narrowing the VFI.
Eliminating the VFI completely may, moreover, not be possible or desirable because of the delicate balance between the level of autonomy necessary to provide incentives for financial control and accountability, and the central government’s ability to insulate itself from demands for sub-national government bailouts through controlled transfers.
The VFI examined in this study is a measure of actual rather than of potential mismatch between revenue and expenditure decentralization. Cyclical developments and one-off revenue and expenditure developments can, however, drive a wedge between actual and potential VFI. Future research could fruitfully examine the impact of these factors on the historical evolution of VFI during periods of economic expansion and decline.
References
Ambrosanio, F., M. Bordignon, and F. Cerniglia, 2008, “Constitutional Reforms, Fiscal Decentralization and Regional Fiscal Flows in Italy,” Working Paper IEF0084 Università Cattolica del Sacro Cuore.
Bach S.H.B., and D. Wallau, 2009, “The Spending Power of Sub-Central Governments: A Pilot Study,” OECD Economics Department Working Paper No. 705, ECO/WKP (2009) 46 (Paris: Organization for Economic Cooperation and Development).
Bahl, R., and S. Wallace, 2007, “Intergovernmental Transfers: The Vertical Sharing Dimension,” in Fiscal Equalization: Challenges in the Design of Intergovernmental Transfers, ed. by J. Martinez-Vazquez and B. Searle (New York: Springer).
Balassone, F., and others, 2008, “Fiscal Consolidation in an Evolving Institutional Framework: the Italian Experience,” Servizio Studi, Ministero dell’Economia e delle Finanze.
Baskaran, T., 2010, “On the Link between Fiscal Decentralization and Public Debt in OECD Countries,” Public Choice, Vol. 145, pp. 351–78.
Bibbee, A., 2007, “Making Federalism Work in Italy,” OECD Economics Department Working Paper No. 590 (Paris: Organization for Economic Cooperation and Development).
Bibbee, A., and F. Padrini, 2006, “Balancing Health Care Quality and Cost Containment: the Case of Norway,” Economics Department Working Paper No. 481 (Paris: Organization for Economic Cooperation and Development).
Bird, R., and A. Tarasov, 2004, “Closing the Gap: Fiscal Imbalances and Intergovernmental Transfers in Developed Federations,” Environment and Planning C: Government and Policy, 2004, Vol. 22, pp. 77–102.
Blöchliger, H., and D. King, 2006, “Fiscal Autonomy of Sub-Central Governments,” OECD Working Paper No. 2. (Paris: Organization for Economic Co-operation and Development).
Blöchliger, H. and O. Petzold, 2009, “Taxes or Grants: What Revenue Source for Sub-Central Governments?” OECD Economics Department Working Papers 706 (Paris: Organization for Economic Cooperation and Development).
Bordignon, M., 2000, “Problems of Soft Budget Constraints in Intergovernmental Relationships: The Case of Italy,” IADB Research Network Working Paper #R-398.
Borge, L.-E., 2006, “Centralized or Decentralized Financing of Local Governments? Consequences for Efficiency and Inequality of Service Provision,” Working Paper No. 14/2006, Department of Economics, Norwegian University of Science and Technology.
Buglione, E. and M. Maré, 2008, “The Taxing Powers of Sub-National Governments and the Role of Own Taxes in Italy: Issues and Perspective,” Presented at the OECD Conference on Financing Sub-Central Governments: Explaining the Balance Between Taxes and Grants.
Buglione, E. and M. Maré, 2010 “Tax Competition and Fiscal Federalism in Italy,” 2010, Presented at the OECD Conference on Tax Competition between Sub-Central Governments.
Charbit, C., “Explaining the Sub-national Tax-Grant Balance in OECD Countries,” WP (2009) 11, OECD, Paris, 2009.
Eyraud, L., and L. Lusinyan, 2011, “Decentralizing Spending More than Revenue: Does It Hurt Fiscal Performance?” IMF Working Paper 11/226 (Washington: International Monetary Fund).
Fattore, G., 1999, “Cost Containment and Reforms in the Italian National Health Service,” in Health Care and Cost Containment in the European Union, ed. by E. Mossialos and J. Le Grand, pp. 733–81.
Fiorino, N., and R. Ricciuti, 2007, “Legislature Size and Government Spending in Italian Regions: Forecasting the Effects of a Reform,” Public Choice, Vol. 131, pp 117–25.
Garcia-Mila, T., and T. McGuire, 2002, “Fiscal decentralization in Spain: An asymmetric transition to democracy,” Economics Working Paper 866, Department of Economics and Business, Universitat Pompeu Fabra, Barcelona.
International Monetary Fund, 2003, Belgium: Selected Issues Paper, IMF Country Report No. 03/50 (Washington).
International Monetary Fund, 2010, Spain: Staff Report for the Article IV Consultations (Washington).
International Monetary Fund, 2010b, “Macro-fiscal Implications of Health Care Reform in Advanced and Emerging Economies,” Available at: http://www.imf.org/external/np/pp/eng/2010/122810.pdf.
Jin, J., and H.-F. Zou, 2001, “How Does Fiscal Decentralization Affect Aggregate, National, and Subnational Government Size?” Journal of Urban Economics, Vol. 52, pp. 270–93.
Joumard, I., and C. Giorno, 2005, “Getting the Most Out of Public Sector Decentralization in Spain,” OECD, Economics Department Working Paper No. 436.
Leibfritz, W., 2009, “Fiscal Federalism in Belgium: Main Challenges and Considerations for Reform,” Economics Department Working Paper No. 473, OECD.
López, G., J. Costa-i-Font, and I. Planas, 2004, “Diversity and Regional Inequalities: Assessing the Outcomes of the Spanish System of Health Care Services,” Working Papers 745, Department of Economics and Business, Universitat Pompeu Fabra, Barcelona.
Magnussen, J. T., P. Hagen, and O. Kaarboe, 2007, “Centralized or Decentralized: A Case Study of Norwegian Hospital Reform,” Social Science and Medicine, 2129–37.
Miaja, M., 2005, “Fiscal Discipline in a Decentralised Administration: The Spanish Experience,” OECD Journal on Budgeting, Vol. 5, No. 2.
Minas, R., 2010, “(Re)centralizing Tendencies Within Health Care Services,” Institute for Futures Studies 2-10:9, Stockholm.
Molero J.-C., 2001, “Analysis of the Decentralization of Public Spending in Spain,” Public Finance and Management, 1 (4), 500–56.
Mora, T., 2009, “Spanish Regional Public Budget Reactions to Revenue Gains: the Dependence on the Transfer System,” Urban Public Economics Review, Núm. 10, 2009, pp. 73–91 Universidad de Santiago de Compostela España.
Piperno, S., 2000, “Fiscal Decentralization in Italy: Some Lessons”, draft presentation at IMF Conference on Fiscal Decentralization.
Rattsø, J., 2003, “Vertical Fiscal Imbalance and Fiscal Behavior in a Welfare State: Norway” in Fiscal Decentralization and the Challenge of Soft Budget Constraint edited by J.A. Rodden, G.S. Eskeland, and J. Litvack, The MIT press.
Rodden, J., 2002, “The Dilemma of Fiscal Federalism: Grants and Fiscal Performance around the World,” American Journal of Political Science, Vol. 46, No. 3, pp. 670–87.
Saltman, R., 2008, “Decentralization, Re-centralization and Future European Health Policy,” European Journal of Public Health, Vol. 18, No. 2, 104–6.
Vinuela, J., 2001, “Fiscal Decentralization in Spain,” draft presentation at IMF Conference on Fiscal Decentralization.
Verdonck, M., P., Cattoir, and K., Algoed, 2009, “Assessment of the Belgian Fiscal Equalization Mechanism and Proposals for Reform,” Actes du Congrès des économistes belges de langue française, CIFoP, Charleroi.
The author is grateful for comments from Benedict Clements, Julio Escolano, Luc Eyraud, Kevin Fletcher, Keiko Honjo, Lusine Lusinyan, Rafael Romeu, Antonio Spilimbergo, and Irina Yakadina. Pierre Jean Albert, Jeffrey Pichocki, and Mileva Radisavljevic provided editorial help.
The tax-sharing arrangement in Belgium differs from that in other OECD countries where sub-federal governments receive a fixed share of actual tax revenues. In Belgium, the total annual amount of the “shared income tax” is derived from a lump sum amount. This amount, which was defined in 1989, is indexed to the CPI and real GDP, and depends on the yield of the PIT in the region (the territoriality principle).
Regions can levy their own taxes on any activities not taxed at the federal level, with some restrictions on the autonomy to set the tax base. The regions also have the right to piggyback onto the federal PIT (although this possibility is not used) as well as to grant PIT reductions. Tax surcharges or reductions do not affect the level of shared taxes; they affect only regional own revenues (Leibfritz, 2009).
In 1993 a municipal property tax was instituted while health contributions were assigned to regions already in 1992. A regional tax on company value added, the IRAP, was introduced in 1998, as well as national income tax surcharges—for all levels. Tax autonomy is higher for ordinary regions than for municipalities, and lower for special regions, where tax revenue mostly comes from sharing of PIT and VAT (Buglione and Maré, 2010).
The new rules awarded stronger role to the mayors and greater independence from political parties which, together with greater budget autonomy, triggered a number of fiscal consolidation reforms (Fiorino and Ricciuti, 2007).
These improvements in revenue performance coincide with the introduction of two important taxes for local governments: the municipal property tax in 1993 and the regional tax on value added at the firm level, in 1998 (Ambrosanio and others, 2008).
Social spending comprises the locally funded safety net including social, minimum and disability pensions; social housing and nurseries; unemployment benefits; etc.
Containment of public spending on health contributed to the consolidation, declining from 6 ½ percent of GDP in 1991 to 5½ percent of GDP in 1999. The measures that contributed to cost containment in the health sector included the control of NHS salaries (1991–95), as well as consolidation of the LHAs that were also granted administrative and financial independence during this period. The reform also introduced quasi-markets in which providers competed for contracts. Later, in 1996, a national program for bed closure was launched and DRGs were introduced in 1994, with the aim of introducing market competition, containing costs, and enhancing efficiency (IMF, 2010b).
In some tax-funded health systems, namely: Norway, Sweden, Finland, Spain, and Italy; administrative, managerial, fiscal, and political decision-making were decentralized between the 1970s and 1990s. The Nordic countries, Germany, and the Netherlands devolved administrative, managerial, and fiscal decisions (but not political ones) to non-profit institutions, while Central European countries fully devolved hospitals to regions and municipalities (Saltman, 2008).
OECD Health Data; and Bibbee and Padrini, 2006.
Law on Hospital Services.
In 2007, the number of RHEs was reduced further from 5 to 4.
Aside from recentralization, fiscal discipline was to be aided by sub-national budget rules. Counties and hospitals were given annual global budgets that had to be balanced ex ante. This rule was not binding, as deficits had to be repaid within the following two years (up to four years in some circumstances). Investments, on which local governments had high discretion, could be financed by central-government-approved loans (Rattsø, 2003).
During 2002–04 central government explicitly adopted an expansionary fiscal policy. Following half a decade of overall budget surpluses and accumulation of oil fund assets, pressures to spend part the oil dividend mounted and guidelines for an accelerated use of oil resources were introduced in March 2001 (IMF, 2002). A very large public sector pay increase was granted in 2002, especially to hospital doctors, and hospital investments increased sharply (Bibbee and Padrini, 2006).
The provinces’ role is limited, their main responsibility being to provide technical and financial assistance to municipalities with fewer than 5,000 inhabitants and invest in certain services. Their tax-raising powers are also limited. They have the power to levy surtax on the business tax in the form of a single rate up to 40 percent.
The foral regime regions levy their own taxes and transfer a subsidy to the central government which is used to finance central government’s public services. The remaining 15 ACs have a common regime status.
The 1997 reforms introduced the tax “co-responsibility principle” allowing ACs to benefit from 30 percent of national income tax receipts. The ACs were given the power to adjust tax rates and tax bases within defined limits.
The 2002 agreement broadened both the range of tax bases and the regions’ discretionary power to set rates. The ACs’ share of personal income tax has increased to 33 percent and the main consumption taxes have been included among the shared taxes, with power to change the rates within a margin.
The share of “ceded” and shared taxes in total revenues for the common regime ACs should reach 90 percent.
Equalization grants were adjusted annually to reflect the central government’s tax revenues, irrespective of each individual region’s actual economic performance. However, to reduce differences in revenues across regions, the system specified maximum and minimum growth rates for overall AC resources (Joumard and Giorno, 2005).
In this case, when revenue decentralization exceeds expenditures responsibilities, the VFI is negative, instead of a vertical fiscal gap, sub-national governments are facing a vertical surplus. A negative VFI can also pose problems: in Italy, the cost of newly assigned functions was unknown, and this was believed to have contributed to overspending and the creation of overlapping responsibilities (Piperno, 2000). The problem was addressed only in the 1990s when further spending competencies were transferred, and specific purpose transfers were cut, thus narrowing the VFI.
