Fiscal Politics
Chapter

Chapter 14. On the Determinants of Fiscal Noncompliance: An Empirical Analysis of Spanish Regions

Author(s):
Vitor Gaspar, Sanjeev Gupta, and Carlos Mulas-Granados
Published Date:
April 2017
Share
  • ShareShare
Show Summary Details
Author(s)
Mar Delgado-Téllez, Victor D. Lledó and Javier J. Pérez 

Introduction

The process of fiscal consolidation in Europe in the aftermath of the global and euro sovereign debt crisis has brought to the forefront the challenges of enforcing fiscal discipline in federal or decentralized countries. The literature on fiscal federalism has attributed this challenge to the presence of soft budget constraints at the subnational level,1 that is, the inability of subnational governments (SNGs’) to keep fiscal deficit outcomes within targets set as part of fiscal consolidation strategies at the general government level. Soft budget constraints have been shown to originate from the inability of central governments (CGs) to credibly commit to not bailing out SNGs and, as a result, to not constrain SNGs fiscal outcomes (Vigneault 2007). Soft budgets have been shown to be driven by political motives, including reelection and government formation and stability (Sato 2007). They are aggravated by flawed intergovernmental fiscal institutions, including large vertical fiscal imbalances, weak fiscal rules, and limited market discipline (Rodden, Eskelund, and Litvack 2003; Ter-Minassian 2015). Flawed institutions act by raising expectations among voters and creditors that the CG must be accountable in the event SNGs are not able to fulfill their spending mandates or debt obligations.2 Soft budget constraints have been typically assessed by exploring the determinants of fiscal outturns using fiscal reaction functions.3

A small but growing empirical literature on the implementation of fiscal consolidations offers a different perspective. Rather than searching for reasons why fiscal outcomes cannot be constrained and targets enforced, it questions whether fiscal targets or the forecasts on which such targets are based are set appropriately in the first place.4 A number of papers have shown that official forecasts tend to be optimistic among advanced economies (Auerbach 1999; Leal and others 2008; Jonung and Larch 2006; Frankel and Schreger 2013). Optimistic fiscal forecasts have been attributed to difficulties in forecasting downturns and booms in real time (Beetsma and others 2013). Another set of factors is related to strategic considerations, which have been shown to be salient in the EU among countries seeking to comply with the Maastricht convergence process (Strauch and von Hagen 2009) and ex ante deficit rules under the Stability and Growth Pact (SGP) (Bruck and Stephan 2006; Beetsma, Giuliodori, and Wierts 2009).

This chapter contributes to both literatures by seeking to provide a better understanding of the determinants of fiscal noncompliance at the subnational level. Fiscal noncompliance is defined as SNG budget balance outturns below corresponding targets. The focus is to understand whether fiscal noncompliance is the result of soft budgets or of technical and institutional factors that result in unrealistic fiscal targets. An emerging empirical literature has begun to look at the determinants of compliance in rules-based frameworks (Cordes and others 2015; Reuter 2015). However, this literature has mostly focused on national policies and has not discussed the institutional and political considerations behind fiscal noncompliance.

This chapter proposes a conceptual framework that tries to distinguish the impact of a soft budget constraint from that of fiscal forecasting and target setting on fiscal noncompliance. The framework looks at both the capacity and the incentives to comply. It distinguishes between events when SNGs have the capacity but not the incentives to comply with fiscal targets from events when SNGs have the incentives but not the capacity for fiscal compliance. Fiscal noncompliance is defined as voluntary under the former and involuntary under the latter. The discussion argues that voluntary fiscal noncompliance is triggered by factors conducive to soft budget constraints, whereas involuntary fiscal noncompliance is the result of factors conducive to unrealistic or ambitious fiscal targets.

Political economy channels and politics take a front seat in this framework. The framework shows that both voluntary and involuntary fiscal noncompliance occur mainly through political economy channels that jointly influence CGs’ and SNGs’ decisions to, respectively, enforce and comply with fiscal targets. Channels conducive to voluntary fiscal noncompliance act mainly by increasing CGs’ political costs of enforcing and decreasing SNGs’ costs of noncomplying with fiscal targets. Channels conducive to involuntary fiscal noncompliance are those that increase CGs’ political cost of ensuring fiscal targets at the general government level are met, leading the CG to shift the burden of meeting these targets to SNGs. Such costs are determined by the impact such decisions have on the electoral, government formation, and other political objectives government officials and their parties have at the central and subnational levels, which is ultimately framed by politics and political institutions at the supranational, national, and regional levels.

An empirical model is constructed to test this framework. From among a set of economic, institutional, and political factors, the model identifies the ones most relevant to an understanding of voluntary and involuntary fiscal noncompliance. The empirical model is estimated using data from Spain’s Autonomous Communities. Spain’s Autonomous Communities (hereafter also referred to as regions, regional governments, or simply RGs) makes for an interesting case study for a number of reasons. RGs have gained significant political and fiscal autonomy over the past four decades through a process of decentralization (Hernández de Cos and Pérez 2013). During this period, regional governments have become accountable for delivering more than two-thirds of social services, mostly in the health and education sectors (Lledó 2015). The Spanish decentralization has been asymmetric, with revenue and expenditure decentralization occurring at different paces depending on the region, leading to both temporal and cross-sectional variations in both fiscal and political autonomy indicators. Spain’s RGs have been subject to nominal budget balance targets for the past two decades. Their record in meeting these targets, as discussed below, has also varied significantly. And so has the rules-based framework used to monitor and enforce compliance with those targets. In addition to fiscal rules, regions have been subject to market-imposed discipline, given that most RGs’ debt is regularly scrutinized by rating agencies. In this respect, Spain is one of the major subsovereign bond issuers worldwide, presenting a significant heterogeneity across regions in issuing practices and amounts (Canuto and Liu 2013; Pérez and Prieto 2015).

The postcrisis period in Spain has been marked by widespread noncompliance. Regions as a group have missed their targets systematically every year since 2010, accounting for the bulk of the fiscal noncompliance at the general government level and constituting one of the main risks to Spain’s ongoing fiscal consolidation process (AIReF 2016). Critical to this analysis, fiscal noncompliance, while widespread, varied significantly across regions in both frequency and margins.

The existing empirical literature has studied fiscal discipline among Spanish regions by assessing the determinants of fiscal deficit and public debt outturns (for example, Argimón and Hernández de Cos 2012; Hernández de Cos and Pérez 2013). This literature has typically looked at economic, institutional, and political factors affecting the size of fiscal outturns irrespective of the targets aimed at constraining them. Critical factors promoting fiscal discipline included greater tax autonomy, higher market-financing costs and credit ratings, and the electoral calendar, but fiscal rules and other political factors are excluded. Fiscal indiscipline appears to have a strong inertial component, with the size of regions’ fiscal deficits in one year largely influenced by the size in the previous year. A related literature has also looked at the determinants of the CG’s budgetary deviations of the CG (Leal and Perez 2011). To the authors’ knowledge, Leal and López Laborda (2015) and Lago-Peñas, Fernández-Leiceaga, and Vaquero (2016) are the only empirical analyses examining the regional determinants of compliance with fiscal deficit targets among Spanish regions.

The rest of this chapter is organized as follows. The next section proposes a conceptual framework to identify economic, institutional, and political determinants of fiscal noncompliance in multilevel governance systems. The third section reviews key institutional elements in Spain’s multilevel governance system, with a focus on how fiscal targets are set, monitored, and enforced. Informed by the framework and Spain’s institutional features, the fourth section proposes alternative hypotheses, details the empirical methodology used to test these hypotheses, and discusses the empirical results. The final section concludes with some policy considerations.

Fiscal Noncompliance in Multilevel Governments: A Conceptual Framework

Defining Fiscal Noncompliance

The proposed framework defines fiscal noncompliance as the outcome when a government is unable to meet numerical fiscal targets or ceilings. The fiscal target or ceiling could be the numerical limit of a fiscal rule. A government unable or unwilling to meet a fiscal target or ceiling is defined as noncompliant.

Fiscal noncompliance can be voluntary or involuntary. Fiscal noncompliance is voluntary when the noncompliant government has the capacity but not the incentives to comply with a fiscal target. Fiscal noncompliance is involuntary when the noncompliant government has the incentives but not the capacity to comply with a fiscal target. A government has the capacity to meet the target if it has sufficient fiscal resources or fiscal instruments to garner the necessary resources to meet the target—hereafter termed fiscal capacity. A government has the incentives to meet the target when the costs of noncomplying with the target outweigh the noncompliance benefits.

The Fiscal Noncompliance Problem

The fiscal noncompliance problem can be characterized as a sequential game between a central and a regional government (panel 1 of Figure 14.1). In the first stage, the CG sets a fiscal target for the RG, knowing the RG’s expected fiscal capacity. The fiscal target is ex ante feasible. In the second stage, the RG decides whether to comply with the fiscal target based on expectations about its fiscal capacity and on whether the CG will enforce the fiscal target. In the third and final stage, the CG decides whether to enforce the target based on the RG’s compliance decision in the second stage and its expected fiscal capacity. Nature reveals itself only at the end of the game in the form of a shock affecting the RG’s fiscal capacity and, therefore, the feasibility of the fiscal target.5

Voluntary and involuntary fiscal noncompliance may also emerge as equilibrium outcomes under this game. Voluntary fiscal noncompliance occurs when the RG is not willing to comply with the budget balance target regardless of whether the CG is expected to enforce it, and even when fiscal capacity to comply with the target is highly expected. Under these circumstances, the shock can be assumed away, because the target is feasible both before and after the shock—that is, the target is both ex ante and ex post feasible—(panel 2 of Figure 14.1). Involuntary fiscal noncompliance occurs when the RG is willing to and ex ante capable of complying, but does not have the ex post fiscal capacity to do so (panel 3 of Figure 14.1).6

Figure 14.1.The Fiscal Noncompliance Problem

Source: Authors’ calculations.

Note: CG = central government; RG = regional government.

Voluntary Fiscal Noncompliance and Soft Budget Constraints

Voluntary fiscal noncompliance could be the result of soft budget constraints. RGs with soft budgets are not constrained to finance their spending from an approved budget. Therefore, they would not feel constrained to deviate from fiscal targets set in this budget if doing so will prevent them from providing a desired level of public goods and services. In the multilevel government context, the soft budget constraint problem arises from the CG’s lack of a credible no-bailout commitment that allows RGs to overspend in the expectation of an eventual bailout.7

Soft budget constraints and voluntary fiscal noncompliance are interconnected. The theoretical literature models soft budget constraints as a sequential game (Inman 2003; Rodden, Eskelund, and Litvack 2003; Vigneault 2007; Bordignon 2006). Actions in the voluntary fiscal noncompliance game described above are logical extensions of the soft-budget-constraint game (Figure 14.2). In the first stage, the CG announces its intergovernmental transfer policy and sets the RG’s budget balance target. In the second stage, the RG does not believe the CG’s transfer policy, expects a bailout, overspends, and thus deviates from the budget balance target. In the third stage, the CG fulfills the RG’s expectation by bailing it out, thereby not enforcing the breach in the budget balance target.8 Much like in the voluntary fiscal noncompliance game, nature’s draw does not make a difference and the target remains feasible.

Figure 14.2.Soft Budget Constraint and Fiscal Noncompliance Problems

Source: Authors’ calculations.

Note: CG = central government; RG = regional government.

Bailout and overspending incentives complement each other to spur voluntary fiscal noncompliance. Two necessary but not sufficient conditions characterize soft budgets and noncompliant governments. The first is that the CG must find it optimal not to enforce the fiscal target and to provide additional resources to the RG in stage 3. It will do so if the economic and political costs of denying additional resources (see below)—thereby enforcing the target—exceed the bailout or nonenforcement costs in the form of administrative, legal, or financial penalties, or if the bailout (nonenforcement) is triggered by deviations from national or supranational fiscal rules as well as reputational losses against financial markets and the public at large. Under these circumstances, the bailout or nonenforcement strategy is ex post optimal. The second necessary condition is that the RG, knowing that the CG has an incentive to provide additional resources and not to enforce the target, finds it optimal to overspend and not comply in stage 2 (that is, overspending is ex ante optimal). An ex post optimal bailout will not lead to noncompliance if overspending is not optimal. This may occur, for instance, if a bailout comes with costly conditions attached (for example, loss of fiscal autonomy, unpopular reforms). At the same time, by construction, an overspending optimal strategy cannot exist in the absence of an ex post optimal bailout. In short, for voluntary fiscal noncompliance to occur, factors that raise both bailout and overspending incentives must be in place.

Bailout and Overspending Incentives

CGs may choose to bail out RGs for economic and political reasons.

  • Economic motives. A benevolent CG that cares for the welfare of the whole nation would choose to bail out a fiscally irresponsible RG to avoid the negative spillovers to other jurisdictions and to itself. Negative spillovers to other jurisdictions—referred to as horizontal spillovers—usually take the form of underprovision of health, education, and other essential services by the non-rescued RG to other RGs. Negative spillovers to the CG, or more broadly, to the general government—referred to as vertical spillovers—may occur if default of a nonrescued RG endangers the banking system or the corporate sector nationwide because of their exposure to RG debt, thereby increasing fiscal risks and lowering credit ratings at the central or general government level (Inman 2003). Bailout incentives are expected to decrease with bailout pecuniary costs for CGs and increase with bailout economic benefits. Pecuniary costs are expected to increase with the size of the region: the larger the region, the larger the cost of the public goods and services it provides. However, the impact of region size on bailout economic benefits is ambiguous and depends on assumptions about the “extensive” and “intensive” nature of the spillover. The larger the region, the larger the extensive nature of the spillover: the larger the number of regions and individuals benefiting from the public goods and services provided by that region, the larger the bailout economic benefits (Wildasin 1997). But the smaller the region, the larger is the intensive nature of the spillover, and the larger the amount of public goods and services appropriated by each citizen in the bailed-out region (Crivelli and Stahl 2013). Bailout incentives are, therefore, expected to increase with RG size if the bailout benefits from the extensive nature of the negative spillovers outweigh both the benefits from its corresponding intensive nature and the bailout pecuniary costs (Wildasin 1997). Otherwise, bailout incentives are expected to decrease with RG size (Crivelli and Stahl 2013).

  • Political motives. CGs may also bail out RGs to create the conditions to govern, stay in power, and reelect their principals. Bailout incentives are greater if directed toward RGs that are well represented in the national legislature, and thus influential for government stability and the passage of critical legislation (Porto and Sanguinetti 2001). Similar motives may also lead CGs to bail out regions with which they are politically aligned—that is, regions where government incumbents are from the same party or coalition of CG incumbents (Grossman 1994).9 The CG may also offer bailouts to ensure national unity (Leite-Monteiro and Sato 2003). As a result, bailout incentives are likely to increase in regions where representation at the national or subnational level of pro-autonomy parties is larger (Bolton and Roland 1997).

Flawed intergovernmental fiscal frameworks increase bailout and overspending incentives. They do so by raising expectations among voters and creditors that the CG must be accountable in the event RGs are not able to fulfill their spending mandates or debt obligations (von Hagen and Eichengreen 1996). Mindful of the political costs of not fulfilling those expectations, CG bailout incentives will likely increase, raising RGs’ bailout expectations and increasing overspending incentives. Rodden, Eskelund, and Litvack (2003) and Ter-Minassian (2015) list a number of institutional flaws that can be broadly categorized as (1) limited fiscal autonomy, (2) lack of preconditions for market discipline, and (3) weak administrative controls and fiscal rules. Limited fiscal autonomy may be the result of RGs’ limited taxing powers, spending discretion limited by minimum service standards or revenue earmarking, and overlapping and unclear revenue or spending assignment. Insufficient fiscal autonomy is usually reflected in large gaps between the RG’s mandated spending and revenue assignments, that is, large vertical fiscal imbalances (VFIs). The capacity of financial markets to discipline RGs is undermined by regulatory incentives and lax prudential requirements on RG lending, RGs’ access to noncompetitive financing sources (CG onlending, public and development banks, state-owned enterprises), and lack of transparent and comprehensive public accounts that blur RGs’ creditworthiness. Administrative controls such as those guiding RG borrowing are usually not based on clear and objective criteria (for example, ability to service debt). Last, fiscal rules applied to RGs are often poorly designed and weakly enforced.

Common-pool financing provides incentives for overspending. When most RG spending is financed out of a common pool of resources with few strings attached, overspending—and by implication noncompliance—will become an attractive option. This will be the case because RGs will bear only a fraction of the marginal costs of providing regional goods and services (von Hagen 2005). Common-pool financing is usually provided in the form of general purpose, open-ended, and equalization transfers or through debt-mutualization schemes. The literature shows that excessive dependency on such transfers to finance subnational public goods and services exacerbates overspending.10

Involuntary Fiscal Noncompliance and Fiscal Stress

Involuntary fiscal noncompliance may become likelier in times of fiscal stress. These are periods marked by large negative fiscal shocks usually associated with significant economic downturns and large fiscal adjustment efforts. In combination, the two factors have been shown to undermine RG capacity to meet fiscal targets as follows:

  • Shocks and forecast errors. Economic shocks commonly trigger fiscal stress, making ex ante feasible targets ex post infeasible. Shocks could be region specific (idiosyncratic shock) or they could affect the whole country (common shock). A common shock can affect regions differently depending on each region’s economic structure (for example, a bust in housing prices would affect regions where preshock median property values had been higher) or exposure to fiscal risks (for example, size of explicit or implicit contingent liabilities assumed by RGs on behalf of public enterprises or regional banks). Large shocks are usually reflected in large forecast errors.11

  • Feasible targets and adjustment plans. In times of fiscal stress, CGs, as guardians of fiscal sustainability, are under pressure from markets and supranational institutions to design and implement ambitious but credible fiscal adjustment plans. Such pressure often leads to ex ante feasible, but very demanding, fiscal targets for the general government (Beetsma, Giuliodori, and Wierts 2009). This is particularly the case for the so-called Stability and Convergence Programs of Europe’s SGP. In such programs, fiscal targets need to show ex ante compliance with SGP fiscal rules. Ambitious but feasible general government targets in decentralized fiscal frameworks are, in turn, often reflected in ambitious but feasible subnational fiscal targets, as CGs try to shift part of the fiscal adjustment effort to regions by “passing the buck” (Vammalle, Allain-Dupre, and Gaillard 2012).12 Involuntary fiscal noncompliance, as a result, is expected to become likelier as fiscal adjustment to meet a given fiscal target increases. RG adjustment efforts, in turn, may increase if fiscal targets are not revised following fiscal noncompliance in a given year, leading to persistent fiscal noncompliance patterns. Similar arguments explain why CG incentives to enforce RG fiscal targets also increase in times of fiscal stress. Failure to do so will increase the likelihood that general government fiscal targets will be breached and that markets and supranational institutions will hold the CG accountable for general government fiscal noncompliance.

The Spanish Fiscal Governance Framework

Numerical fiscal targets at the regional level go back more than two decades in Spain. They were subject to numerous changes before and after the global financial crisis:

  • Budget consolidation scenarios and the 2002 Budget Stability Law. Regions were first subject to budget balance limits in the form of fiscal deficit ceilings as part of the Budget Consolidation Scenarios agreed to with the CG after 1992. Fiscal deficit ceilings at the regional level came into law four years later under the 2002 Budget Stability Law (BSL). The 2002 BSL set a single zero-deficit limit for all regions, that is, all regions were obliged to post a budget outturn that was either in balance or in surplus. It also envisaged an adjustment plan with corrective actions in the event of noncompliance. Throughout this period, fiscal deficit ceilings for each region were set as a percentage of national GDP.

  • The 2006 Budget Stability Law. The reform of the first BSL, approved in 2006, entered in force in 2007 and was implemented as a consequence of an EU-wide reform of the SGP. The 2006 BSL enabled the CG and RGs to adapt their deficit and surplus targets to the economy’s cyclical position. Specifically, it allowed the RGs to run a deficit of 0.75 percent of GDP if economic growth was below a certain threshold, to which a further 0.25 percent of GDP could be added to finance increases in productive investment.13 Fiscal deficit ceilings were also set as a percentage of regional rather than national GDP. The 2006 BSL included a no-bailout clause. It also introduced monitoring and enforcement mechanisms. If a risk of noncompliance was detected by the Ministry of Finance, a warning could be made to the responsible government unit. In the event noncompliance materialized, the noncompliant government was required to draw up an economic and financial rebalancing plan over a maximum term of three years. Last, it stipulated that, if a deviation from targets were to prompt a breach of the SGP, the tier of government involved should assume the attendant proportion of the responsibilities that should arise from the breach. In addition, RGs that failed to meet the deficit target would require CG authorization to initiate any debt operations.

  • The 2012 Budget Stability Law. Regional fiscal targets were subject to further refinements to comply with EU-wide fiscal governance taking place in the context of the Six Pack, Fiscal Compact, and Two Pack. A constitutional reform approved in 2011 enshrined the rules-based framework in the Constitution. A new BSL approved in 2012 introduced structural budget balance, expenditure, and debt rules at the regional level. The 2012 BSL refined rules-based monitoring and enforcement mechanisms to prevent, correct, and penalize deviations from fiscal rules and targets introduced in the 2006 BSL. Monitoring and enforcement were also reinforced through improvements in the quality, coverage, and frequency of intrayear regional and local budget figures and the creation in 2013 of Spain’s independent fiscal council—Autoridad Independiente de Responsabilidad Fiscal. Fiscal deficit limits continued to be measured as a percentage of regional GDP.

Understanding Fiscal Noncompliance Among Spain’s Regions

Empirical Methodology

Alternative drivers of fiscal noncompliance among Spanish regions are assessed by looking at noncompliance frequencies and compliance margins. To gather some stylized facts, the analysis starts by examining noncompliance empirical distributions across a number of different potential determinants of voluntary and involuntary fiscal noncompliance. An econometric analysis is then performed to identify whether fiscal noncompliance is likely to be voluntary by looking at the determinants of compliance margins. The sample includes 16 out of 17 Spanish regions over the period 2002–15.14

Noncompliance events are defined as cases of negative deviations between fiscal outturns and fiscal targets for a given region and year. That is, fitfit*<0, where f, f*, i, and t are fiscal balance outturns, fiscal balance targets, respectively. Noncompliance events are sourced from the annual compliance report submitted by the Ministry of Finance to the Economic and Financial Council (CPFF).15 The CPFF comprises the Minister of Finance and public finance authorities from each region. While the Ministry of Finance is the ultimate body in charge of overseeing regional finances, the CPFF plays a formal role in the approval of regions’ fiscal balance targets.

Noncompliance frequencies are defined in equation (14.1) as the ratio of noncompliance cases to the total number of cases within a particular group X. Groups are partitioned by quartiles (q) if measured on the basis of a continuous variable.

Compliance margins, fe = ff, are measured in percentages of regional GDP. Officially, they were measured as differences between fiscal outturns and targets as a percentage of national GDP between 2003 and 2007 and as a percentage of regional GDP from 2008 onward. To allow compliance margins to be compared over the years and across regions according to a homogeneous metric that at the same time reflects differences in regions’ fiscal capacity, official compliance margins have been reestimated in percentage of regional GDP using the latest nominal GDP series.16 That was accomplished in two steps: first, nominal deficit values were uncovered by multiplying targets and outturns by the nominal GDP available around the time targets and outturns were, respectively, set and assessed; second, the difference between nominal deficit outturns and targets was divided by the latest nominal regional GDP series.

A dynamic panel regression analysis is used to examine potential determinants of noncompliance margins. Noncompliance margins are regressed on the same variables conditioning noncompliance frequencies. Estimates are derived using Arellano-Bond first-difference generalized method of moments (FD-GMM) estimators to allow for possible inertial patterns in noncompliance as well as endogeneity of dependent variables. Equation 14.2 summarizes the specification:

where INVOL and VOL are vectors with factors associated with involuntary and voluntary noncompliance events (hereafter referred to as voluntary and involuntary factors), respectively; η and ρ are, respectively, country and time fixed effects, a governs the degree of persistence of RG fiscal compliance and noncompliance, and γ and δ measure the relative contribution of involuntary and voluntary factors to fiscal compliance and noncompliance.17

The estimation strategy aims to identify operative economic, institutional, and political factors associated with voluntary and involuntary patterns of fiscal non-compliance. In light of the relatively short cross-sectional dimension, the identification strategy is implemented in a parsimonious way by individually assessing the impact of a larger set of variables expected to encourage voluntary fiscal non-compliance on a baseline that controls for lagged fiscal noncompliance and the more limited number of factors associated with involuntary compliance patterns. To address the problem of overfitting and biased estimates in small cross-section samples stemming from the proliferation of GMM instruments, only the lags t−2 and t−3 are used, and the instruments are combined into smaller sets by using the collapse option in Roodman’s xtabond2 package for Stata. The robustness of the results are checked using two-stage least squares (2SLS) estimators.

Testable Hypotheses

The proposed multilevel governance framework developed in the second section can help us understand fiscal noncompliance among Spain’s regions. It can do so by helping identify to what extent regional fiscal noncompliance is voluntary. Voluntary fiscal noncompliance can be the result of bailout or overspending incentives driven by welfare or political motives. The framework can also look at the role political, fiscal, and financial market institutions play in shaping such incentives. Fiscal noncompliance could have also been involuntary because of common or asymmetric shocks, and because of fiscal targets and adjustment plans that were borderline feasible. Drawing from this framework and empirical analysis referenced in the previous section, Table 14.1 summarizes some testable hypothesis that are relevant in the Spanish context.

Table 14.1.Fiscal Noncompliance Testable Hypotheses
Expected Sign
ChannelVariableNoncompliance FrequencyCompliance Margin
I. Voluntary
SpilloversRegion sizeNegative/positivePositive/negative
Fiscal AutonomyTax autonomyNegativePositive
Expenditure discretionNegativePositive
Market DisciplineFinancing costsNegativePositive
Access to soft financingPositiveNegative
Fiscal RulesFiscal rule strengthNegativePositive
Political RepresentationSize of parliament representationPositiveNegative
Congruence of regional and national government coalitionsPositiveNegative
ElectionsElection yearPositiveNegative
Political AutonomyRegional representation of pro-autonomy partiesPositiveNegative
II. Involuntary
ShocksCommon or nationwide positive shocksNegativePositive
Region-specific positive shocksNegativePositive
Fiscal Target AdjustmentAnnual changes in fiscal deficit targetsPositiveNegative
III. Other
InertiaCompliance margin in previous yearPositive
Source: Authors’ calculations.Note: See Annex 14.1 for a detailed description of the variables. — = not applicable.
Source: Authors’ calculations.Note: See Annex 14.1 for a detailed description of the variables. — = not applicable.

Facts and Factors

Fiscal noncompliance between 2003 and 2015 varied markedly across regions in both how frequently regions missed the target and by how much these targets were missed (Figure 14.3). Fiscal noncompliance frequencies appear to be stratified into at least three groups: (1) broadly compliers, (2) broadly noncompliers, and (3) largely noncompliers. The broadly compliers comprise regions that have stuck to their fiscal targets in at least half of the years during the analysis periods. This is a large and heterogeneous group demographically, economically, and historically. It includes the Canary Islands, Galicia, Madrid, Asturias, Castilla and León, Extremadura, Andalucía, Aragón, and the Basque Country. Navarra, Rioja, Castilla la Mancha, the Balearic Islands, Cantabria, and Murcia are among the broadly noncompliers—regions missing their targets in up to two-thirds of the years. Finally, Valencia and Catalonia missed their fiscal targets in three-fourths of the years during this period. Just like the first group, regions in the last two groups have very distinct attributes. Noncompliance frequencies and margins appear to be broadly correlated in the sense that more-frequent noncompliers tend to breach their targets by wider margins than less-frequent ones.

Figure 14.3.Regions’ Noncompliance with Fiscal Deficit Targets

Sources: Ministry of Finance; and authors’ calculations.

Note: Under the official assessment, fiscal noncompliance events are defined as differences between fiscal targets and outturns in percentage of national GDP between 2003 and 2007 and as a percentage of regional GDP from 2008 to 2015. Under the homogeneous assessment, fiscal noncompliance events are defined as differences between fiscal targets and outturns in percentage of regional GDP between 2003 and 2015 (see Annex 14.1). AND = Andalusia; ARA = Aragon; AST = Asturias; BAL = Balearic Island; CAN = Canary Islands; CANT = Cantabria; CAT = Catalonia; CL = Castilla and Leon; CM = Castilla La Mancha; EXT = Extremadura; GAL = Galicia; MAD = Madrid; MUR = Murcia; NAV = Navarra; PV = Basque Country; RIO = Rioja; VAL = Valencia.

Regions’ fiscal noncompliance increased markedly in the postcrisis years. The number of noncompliant regions and their corresponding noncompliance margins also increased significantly following the global financial crisis (Figure 14.4). Noncompliance peaked in the post–EU sovereign debt crisis in 2011 when virtually all regions were unable to meet their fiscal deficit targets, most of them by very large margins. This deviation was corrected in the following years through more realistic projections of shared revenues advanced to the regions and supported by fiscal adjustment plans.

Figure 14.4.Evolution of Regions’ Noncompliance with Fiscal Deficit Targets

Sources: Ministry of Finance; and authors’ calculations.

Note: Under the official assessment, fiscal noncompliance events are defined as differences between fiscal targets and outturns in percentage of national GDP between 2003 and 2007 and as a percentage of regional GDP from 2008 to 2015. Under the homogeneous assessment, fiscal noncompliance events are defined as differences between fiscal targets and outturns in percentage of regional GDP between 2003 and 2015 (see Annex 14.1).

Involuntary Channels and Baseline Specifications

Fiscal noncompliance, common shocks, and forecast errors are linked. Common shocks are proxied by observed deviations between nominal (national) GDP growth outturns and forecasts set in annual budget laws (forecast errors).18 Negative (positive) forecast errors in nominal GDP growth should undermine (bolster) compliance with fiscal deficit targets through corresponding revenue shocks. Noncompliance margins and frequencies have clearly moved in tandem with forecast errors (Figure 14.5). Years when fiscal noncompliance was widespread (2008–11 and 2014–15) have usually been years when forecast errors have been negative.19 Regression results provide support for the positive correlations between forecast errors and involuntary fiscal compliance, with positive and statistically significant estimates in about half of all estimated models (Tables 14.2 and 14.3).

Figure 14.5.Forecast Errors and Regions’ Noncompliance with Fiscal Targets

Sources: Ministry of Finance; and authors’ calculations.

Idiosyncratic shocks seem to play a limited role in determining fiscal noncompliance. Measured by differences between regions’ real GDP growth, consumer price inflation, and house price inflation, and corresponding national averages, positive idiosyncratic shocks are expected to reduce fiscal noncompliance frequencies (Figure 14.6). Contrary to expectations, noncompliance frequencies were either the same (real GDP growth) or larger (consumer price inflation and house inflation) among cases in which idiosyncratic shocks were positive. Equally unexpected, positive idiosyncratic growth shocks seem to reduce rather than increase fiscal compliance margins. However, country-specific inflation differentials are not shown to be statistically significant (Tables 14.2 and 14.3). As discussed below, this finding may be explained by the relatively strong transfer dependency observed in most regions and, more specifically, by the fact that a significant share of regional finances comes in the form of transfers from the center allocated with the objective of equalizing regions’ fiscal capacity to meet their spending mandates. Thus, reliance on equalization transfers mitigates the revenue impact of region-specific shocks, helping regions safeguard their fiscal capacity and meet their fiscal deficit targets.

Table 14.2.First-Difference GMM Estimates of Fiscal Compliance Margins
(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)
Lagged Noncompliance Margin0.74*1.09**0.76*0.83*0.760.70***0.370.91**2.42***0.31*0.73*0.490.24
Growth Forecast Errors0.09*0.040.090.10*0.10*0.12***0.040.090.060.050.090.12***0.04
Regional-National Growth Differential–2.09*–2.67*–2.16*–2.13–2–0.73–0.6–2.32*–1.82*–0.36–2.17*–1.190.24
Regional-National Inflation Differential–0.36–2.08–0.39–1.1–1.28–1.27–2.41**–1.23–1.68–2.83*0.86–1.29–1.19
Fiscal Target Adjustment–0.80**–1.13**–0.81**–0.94**–0.85*–0.35–0.82***–1.02**–0.99***–0.51***–0.68*–0.52–0.49
Execution Minus Budgetary Transfers (in regional GDP)–0.48
Regional Weight in National Population3.86
Regional Weight in National GDP7.12
Regional Weight in National Per Capita GDP0.36
Tax Autonomy0.03
Social Spending Share in Regional Government Spending–0.33**
Investment Share in Total Regional Spending–0.24***
Vertical Fiscal Imbalances–0.15***–0.13**
Fiscal Rule Index0.050.020.01
Fiscal Rule Index × Lagged Noncompliance Margin–0.06**
Lagged Annual Change in Region Ratings–0.09
Lagged Annual Change in Implicit Interest Rates1.03***0.19
Ratio of Security to Loans0.00
National Election Dummy–0.60***
Regional Election Dummy–0.36*–0.50*
Party Congruence Dummy–0.180.40–0.41***
Proautonomy Party Share–0.030.02
Regions’ Seats in National Parliament–0.31–0.86
Number of Observations176160176176176160160160160145176176144
Number of Regions16161616161616161615161616
Number of Instruments16151616161515151515171714
Hansen0.600.800.440.870.660.280.370.820.650.140.680.150.05
m10.120.110.130.150.230.020.090.100.060.140.120.200.00
m20.390.610.430.490.420.480.610.660.770.160.280.220.67
Source: Authors’ calculations.Note: The dependent variable is the difference between regions’ fiscal deficit outturns and fiscal deficit targets. The larger this difference, the larger the fiscal compliance margin. Instrument set in all models includes the second and third lag of the explanatory variables. Hansen is the p-value of the test of the overidentifying restrictions (see Hansen 1982), which is asymptotically distributed chi square under the null hypothesis that these moment conditions are valid. A p-value equal to or greater than 0.05 indicates that the instrument set is valid, which is confirmed under all models. Note that m1 and m2 are the p-values of serial correlation tests of order 1 and 2, respectively, using residuals in first differences. The null hypothesis under both m1 and m2 tests is that there is no correlation between variables in the instrument set and the residuals. Observed p-values greater than 0.05 under the m2 test for all models indicate that there is no correlation with the instrument set defined in second lags. GMM = generalized method of moments.*p < .1; **p < .05; ***p < .01.
Source: Authors’ calculations.Note: The dependent variable is the difference between regions’ fiscal deficit outturns and fiscal deficit targets. The larger this difference, the larger the fiscal compliance margin. Instrument set in all models includes the second and third lag of the explanatory variables. Hansen is the p-value of the test of the overidentifying restrictions (see Hansen 1982), which is asymptotically distributed chi square under the null hypothesis that these moment conditions are valid. A p-value equal to or greater than 0.05 indicates that the instrument set is valid, which is confirmed under all models. Note that m1 and m2 are the p-values of serial correlation tests of order 1 and 2, respectively, using residuals in first differences. The null hypothesis under both m1 and m2 tests is that there is no correlation between variables in the instrument set and the residuals. Observed p-values greater than 0.05 under the m2 test for all models indicate that there is no correlation with the instrument set defined in second lags. GMM = generalized method of moments.*p < .1; **p < .05; ***p < .01.
Table 14.3.Two-Stage Least Square Estimates of Fiscal Compliance Margins
(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)
Lagged Noncompliance Margin0.49**0.57**0.49**0.41**0.39**0.52**0.75**0.49**0.460.140.30*0.42**0.79*
Growth Forecast Errors0.04*0.010.04*0.05***0.05**0.03–0.030.030.05*0.07**0.04**0.05**–0.02
Regional-National Growth Differential–0.42**–0.46**–0.42**–0.44***–0.46***–0.42**–0.41*–0.39**–0.12–0.26–0.27**–0.38**–0.40*
Regional-National Inflation Differential–0.47–0.50–0.45–0.54–0.55–0.28–0.50–0.46–0.17–0.440.00–0.51–0.55
Fiscal Target Adjustment–0.71***–0.77***–0.71***–0.66***–0.66***–0.62***–1.05***–0.78***–0.44***–0.53***–0.48***–0.62***–1.04***
Execution Minus Budgetary Transfers (in regional GDP)–0.22
Regional Weight in National Population–2.59
Regional Weight in National GDP2.73***
Regional Weight in National per Capita GDP0.21***
Tax Autonomy0.02
Social Spending Share in Regional Government Spending–0.29***
Investment Share in Total Regional Spending–0.06***
Vertical Fiscal Imbalances–0.16***–0.15***
Fiscal Rule Index0.14***0.040.07*
Fiscal Rule Index × Lagged Noncompliance Margin–0.02
Lagged Annual Change in Region Ratings–0.15
Lagged Annual Change in Implicit Interest Rates0.26***–0.07
Ratio of Security to Loans0.00***
National Election Dummy–0.64***
Regional Election Dummy–0.2–0.27*
Party Congruence Dummy0.040.13–0.19
Proautonomy Party Share–0.00–0.00
Regions’ Seats in National Parliament–0.58**–0.78*
Number of Observations144128144144144128128128128131144144128
Number of Regions16161616161616161616161616
Hausman Test for Endogeneity21.2724.2121.115.9912.7310.2815.9712.640.011.69.7712.6812.7
(0.00)(0.00)(0.00)(0.00)(0.00)(0.01)(0.00)(0.00)(0.93)(0.23)(0.01)(0.00)(0.00)
Source: Authors’ calculations.Note: The dependent variable is the difference between regions’ fiscal deficit outturns and fiscal deficit targets. The larger this difference, the larger the fiscal noncompliance margin. All variables are defined in level differences. The instrument set in all models includes the second and third lag of the explanatory variables. Standard errors allow for correlation within regions but not among regions (cluster robustness specification).*p < .1; **p < .05; ***p < .01.
Source: Authors’ calculations.Note: The dependent variable is the difference between regions’ fiscal deficit outturns and fiscal deficit targets. The larger this difference, the larger the fiscal noncompliance margin. All variables are defined in level differences. The instrument set in all models includes the second and third lag of the explanatory variables. Standard errors allow for correlation within regions but not among regions (cluster robustness specification).*p < .1; **p < .05; ***p < .01.

Figure 14.6.Fiscal Noncompliance and Regions’ Idiosyncratic Exposure to Shocks

(Frequency of noncompliant cases over 2003–15, percent)

Sources: Ministry of Finance; and authors’ calculations.

Fiscal noncompliance has displayed some inertial patterns. In line with Leal and López-Laborda (2015) and Lago Peñas, Fernández-Leiceaga, and Vaquero (2016), fiscal compliance margins appear to be positively autocorrelated (Figure 14.7). As mentioned by Argimón and Hernández de Cos (2012), this could reflect budget rigidities due to incremental budget processes or multiyear expenditure commitments. Tables 14.2 and 14.3 confirm such inertial patterns under several specifications.

Figure 14.7.Inertia in Regions’ Noncompliance with Fiscal Targets, 2003–15

Sources: Ministry of Finance; and authors’ calculations.

Note: Nominal GDP growth forecasts are set in the budget law.

Fiscal noncompliance increases with the required adjustment effort. Adjustment effort is measured by the difference between the fiscal deficit target in years t and t−1, both in percentage of regional GDP, a simple proxy for the required nominal adjustment.20 Adjustment efforts have been quite heterogeneous across regions given that fiscal deficit targets, despite the existence of different starting fiscal positions, have been set uniformly across regions in most years. As expected, adjustment efforts are found to have a negative and statistically significant impact on fiscal compliance margins in most specifications (Tables 14.2 and 14.3). Estimated coefficients range from 0.5 to 1.0, implying that for each percentage point increase in RGs’ fiscal deficit targets, compliance margins should be expected to decline between 0.5 and 1.0 percentage point.

Fiscal noncompliance may decrease if regions benefit from gap-filling transfers before the assessment date, as discussed in the second section. To verify that, we look at differences between actual transfers received by the RG from the CG and those originally budgeted. Noncompliance margins for an RG that receives more transfers than budgeted should be smaller. This hypothesis is rejected, with regression estimates not significant and with the wrong sign (Tables 14.2 and 14.3, model 2). One interpretation is that, while improving regions’ fiscal capacity and thus helping stave off involuntary fiscal noncompliance, additional unbudgeted transfers reinforce expectations of further gap-filling transfers by end-year, thus boosting voluntary fiscal noncompliance and more than outweighing the initial deterrent effect.

Voluntary Channels

The analysis finds some tentative evidence of a positive impact of region size on fiscal noncompliance. Region size is measured according to the weight of a region’s population, GDP, and GDP per capita in the corresponding national figures. Fiscal noncompliance tends to be more frequent among larger regions (that is, toward the end of the distribution) in all three measures, particularly with respect to GDP per capita (Figure 14.8). Fiscal compliance margins are shown to increase in a statistically significant way with regional GDP and regional GDP per capita only under 2SLS models (Tables 14.2 and 14.3, models 3 to 5).

Insufficient fiscal autonomy to adjust seems to play a role in determining regions’ fiscal noncompliance. To assess the impact of fiscal autonomy, measures of tax and expenditure autonomy as well as of VFIs are estimated. Tax autonomy (in line with the terminology in the local public finance literature) is defined as the share of an RG’s total tax revenues over which the RG has some degree of regulatory autonomy.21 The larger this share, the greater a region’s tax autonomy or fiscal coresponsibility, as it is often referred to in the Spanish empirical literature. However, in contrast with the local public finance literature, expenditure autonomy is defined here by the degree of discretion over mandated expenditures. With health and education mostly mandated to regions under center-imposed minimum standards, and social protection shared with the center, a larger share of regions’ spending on these basic services limits regions’ ability to adjust and comply with fiscal targets once their revenue-raising capacity is taken into account. That is, a region’s autonomy to cut expenditures is expected to decrease as that region’s spending share in basic services increases. With that in mind, the shares of regions’ spending on essential public services (health, education, and social protection) and public investment in their total spending are computed.22 Last, following Eyraud and Lusinyan (2013), VFI indicators are estimated for each region to capture the extent to which regions are unable to finance their own spending with own revenues, regardless of whether they have regulatory power of the corresponding tax bases.23 As expected, noncompliance frequencies tend to be smaller among regions in the top tax autonomy quartiles (Figure 14.9), although the relation is not significant with respect to fiscal compliance margins (Tables 14.2 and 14.3, model 6). On the other hand, fiscal noncompliance frequencies are not necessarily the largest among regions in the top expenditure autonomy and VFI quartiles (that is, regions with greater social mandates and less own resources to fund them).24 That said, as expected, fiscal compliance margins decrease as a larger share of regions’ expenditures is allocated to social services and public investment—that is, as regions’ expenditure autonomy decreases (Tables 14.2 and 14.3, model 6). Finally, regions with large VFIs tend to display lower compliance margins, as shown in Tables 14.2 and 14.3 (models 7 and 13).

Figure 14.8.Regions’ Size and Noncompliance with Fiscal Deficit Targets

(Frequency of noncompliant cases over 2003–15, by quartile, percent)

Sources: Ministry of Finance; Ministry of Public Works and Transport; National Institute of Statistics; and authors’ calculations.

The impact of stronger rules on fiscal compliance is not clear-cut. As described in the previous section, fiscal rules in Spain have become increasingly stronger over the years. They are currently among the strictest fiscal rules in Europe, as measured by the European Commission fiscal rule strength index. Stronger rules, however, have not always led to improvements in fiscal compliance, partly because of delays in implementing existing monitoring and enforcement procedures (Lledó 2015). The regression results seem to reinforce this point. Under the baseline GMM specification, stronger fiscal rules do not show any direct impact on fiscal compliance margins directly. Instead, they seem to have an indirect impact on compliance margins by helping reduce inertial patterns (Table 14.2, models 8 and 9). These results are reversed under the 2SLS specification, which shows fiscal rules having a direct rather than an indirect impact on fiscal compliance margins (Table 14.3, models 8, 9, and 13).

Figure 14.9.Regions’ Fiscal Autonomy and Noncompliance with Fiscal Deficit Targets

(Frequency of noncompliant cases over 2003–14, by quartile, percent)

Sources: Ministry of Finance; and authors’ calculations.

Note: Tax autonomy is defined as the ratio between a regional government’s own tax revenue and total tax revenues. Expenditure autonomy is the region’s share of general government spending on essential services. VFI = vertical fiscal imbalance.

Financial markets seem to affect fiscal noncompliance through two different channels. On the one hand, fiscal noncompliance frequencies are larger among regions with lower (poorer) credit ratings and, to some extent, facing larger market-financing costs, which seems to provide some support to the idea that financial markets undermine fiscal compliance by raising the financing costs of regions that are not perceived as creditworthy (Figure 14.10).25 On the other hand, fiscal noncompliance becomes less prevalent among regions where reliance on market-issued securities, rather than softer bank loans, is greater. This finding indicates that greater market exposure helps deter fiscal noncompliance because regions internalize the impact fiscal noncompliance would have on credit ratings and market-financing costs. The regression analysis of fiscal noncompliance corroborates the latter channel: increases in the financing costs faced by regions in the previous year tend to increase rather than reduce compliance margins in the following year (Tables 14.2 and 14.3, model 10). However, greater reliance on market securities has no statistically or economically significant impact on compliance margins (Tables 14.2 and 14.3, model 10).

Figure 14.10.Financial Markets and Regions’ Noncompliance with Fiscal Targets

(Frequency of noncompliant cases over 2003–15, by quartile, percent)

Sources: Ministry of Finance; and authors’ calculations. ‘Regional governments’ credit ratings.

1Regional governments’ credit ratings.

2Ratio of region’s public debt in government securities to banking loans, percent.

3Region’s interest payments in percentage of end-of-year regional public debt stock.

Fiscal compliance is weakened during election years, but the role played by politics in other areas is less clear-cut. Fiscal noncompliance seems to increase during election years. As expected, fiscal noncompliance is more frequent and displays wider margins during election years (Figure 14.11; Tables 14.2 and 14.3, models 11, 12, and 13). Unlike previous fiscal discipline analyses for Spain, but as expected in this framework, political alignment or party congruence between central and regional governments notably increases the likelihood of fiscal non-compliance. In particular, regions politically aligned with the center are shown to be nearly 1.5 times more likely to deviate from targets than nonaligned regions.26 The regression results provide only tentative support to these hypotheses: regions aligned with the center presented smaller, albeit statistically insignificant, compliance margins under most specifications (Figure 14.11; Tables 14.2 and 14.3, models 11, 12, and 13). Pro-autonomy regions, defined by the percentage of members of parliament from regional pro-autonomy parties—expected to deviate from center-imposed fiscal targets—turned out to be only marginally more likely to deviate from fiscal targets than regions with weaker pro-autonomy preferences, with pro-autonomy regions presenting smaller but statistically insignificant margins under most specifications (Figure 14.11; Tables 14.2 and 14.3, models 11, 12, and 13). Last, regions with the largest political representation in the national parliament are the most frequent noncompliers, albeit not necessarily with compliance margins that are statistically significantly smaller (Figure 14.11; Tables 14.2 and 14.3, models 11, 12, and 13).

Figure 14.11.Politics and Regions’ Noncompliance with Fiscal Deficit Targets

(Percent)

Sources: Ministry of Finance; and authors’ calculations.

1‘Regional government led by the same party or by a government coalition leading the central government.

2Percentage of members of regional parliaments from regional pro-autonomy parties.

3Seats in national parliament (lower house) allocated to each region.

Conclusions and Policy Discussion

This chapter argues that in multilevel governance systems, SNGs tend not to comply voluntarily with fiscal targets the larger are their compliance costs as well as the costs the CG is expected to incur in enforcing these targets. It proposes a conceptual framework in which these costs can be, first, political and thus determined by factors directly undermining CGs’ condition to be elected and form stable government coalitions (for example, the national or regional electoral calendar and RGs’ political representation, affiliation, and political autonomy preferences). Second, compliance and enforcement costs are also linked to intergovernmental fiscal frameworks—fiscal rules, tax and expenditure assignments, borrowing controls—and, more specifically, to how these arrangements shape perceptions among voters, creditors, and politicians of SNGs’ fiscal autonomy and whether they rather than CGs should be held politically accountable for any disruption in regions’ fiscal obligations in the event of noncompliance. Lack of fiscal autonomy shifts political accountability to CGs—thereby raising enforcement costs—while stronger rules and access to financial markets tip the political barometer toward RGs—thereby raising noncompliance costs.

In this chapter’s framework, involuntary fiscal noncompliance occurs when SNGs are unable to be fiscally compliant even when they are willing to be. This pattern becomes more likely in times of fiscal stress, defined as periods with large negative fiscal shocks. Fiscal stress times are also periods of increasing domestic or supranational political pressure on CGs to ensure that fiscal consolidation targets at the general government level are met. To minimize the political costs such pressures entail, CGs tend to “pass the buck” of the adjustment down to RGs. This leads to ambitious but feasible center-imposed SNG fiscal targets that become infeasible once the fiscal shock materializes.

Applied to Spain’s regions, this conceptual framework shows that fiscal non-compliance displays involuntary traits. We find fiscal noncompliance to be driven by factors partly outside the control of Spanish regions, namely common macro-economic shocks and large adjustment efforts. The latter is arguably attributable to ambitious and rigid fiscal targets set by the center as a result of national and supranational pressures for general government consolidation referred to above.

Fiscal noncompliance among Spain’s regions has also been shown to have a voluntary dimension, with fiscal rather than political arrangements playing a somewhat more prominent role. Fiscal deficit targets were missed more frequently and by wider margins the lower a region’s autonomy to cut spending due to expenditure mandates and the larger the gap between the resources they can raise to deliver these mandates and their actual costs (that is, the larger VFIs are). Contrary to expectations, stronger and well-enforced fiscal rules have not made

fiscal compliance more frequent or compliance margins wider. The analysis has also identified some tentative support for the disciplinary role of financial markets, with increases in regions’ market-financing costs reducing fiscal noncompliance margins. The frequency and margins of fiscal noncompliance have also been shown to increase during election years. Other political factors expected to induce voluntary fiscal noncompliance, such as political autonomy preferences, political alignment with the center, and political representation, demonstrate ambiguous or nonsignificant regression estimates.

The main policy lesson in this analysis is that enhancing fiscal compliance in multilevel governance systems requires a more comprehensive assessment of intergovernmental fiscal arrangements that goes beyond strengthening formal rules-monitoring and enforcement procedures. This assessment should include not only rules-based fiscal frameworks but also (1) the assignment of revenue-raising and spending mandates and (2) the burden-sharing of fiscal consolidation efforts and related setting of fiscal deficit targets. All that should be accompanied by a focus on making CG enforcement politically credible. In particular,

  • Rules-based frameworks. To strengthen fiscal compliance at the national level, much emphasis has been placed on the need to bolster rules-based fiscal frameworks with formal enforcement procedures such as financial and administrative sanctions and automatic mechanisms that correct for past deviations from fiscal targets (Schaechter and others 2012). That has been the case in Spain, particularly after the most recent reforms, which, as discussed, introduced some of these procedures aimed at tackling regional fiscal noncompliance. Looking ahead, there is still some scope to further strengthen existing procedures by making their activation more automatic and by tightening the legal requirements to publicly explain deviations from fiscal targets (Lledó 2015). Such measures may come in particularly handy during election years when the political costs for the CG to enforce targets are more salient and noncompliance has been shown to be more pervasive than in nonelection years.

  • Intergovernmental fiscal responsibilities. In line with previous work looking at the effectiveness of subnational fiscal rules (Kotia and Lledó 2016), this analysis stresses the need to revisit, and possibly reduce, existing VFIs by ensuring SNGs’ revenue-raising and borrowing mandates are consistent with their spending mandates. These measures would help strengthen SNG fiscal autonomy and policy accountability, including for fiscal deficit targets. In doing so, it would make CG enforcement of SNG fiscal deficit targets politically less costly and more credible.

  • Fiscal consolidation burden sharing. The negative impact of increases in fiscal targets on compliance margins warrants a review of how the burden of fiscal consolidation is shared across and within government levels and, correspondingly, how realistically fiscal deficit targets are set. SNG reputational costs for noncompliance with fiscal targets that are widely perceived as infeasible among voters, markets, and politicians are minimal, rendering even well-designed and well-implemented enforcement mechanisms toothless. In the case of Spain, this may call for the adoption of differentiated fiscal targets across regions to balance adjustment needs with existing fiscal capacity. In light of the impact of negative growth shocks on fiscal compliance, a review is also warranted of how appropriate the technical capacity and procedures behind the formulation of macroeconomic forecasts informing central and subnational budgets and fiscal plans are.

Two additional qualifications are worth mentioning with regard to the normative proposals outlined above that go beyond the scope of this chapter:

  • First, while the adoption of differentiated fiscal targets might be efficient when conditioning on a given fiscal starting position (that is, a given level of regional deficit and debt), in a more general, dynamic setting, moral hazard arguments dictate that SNGs may develop incentives not to conduct sound fiscal policies in good times. This might be the case when SNGs anticipate that additional room for fiscal maneuver is to be granted in crisis times to those governments with weaker initial fiscal positions. The strict implementation of fiscal rules is crucial for the development of ex ante fiscal margins against adverse shocks, and to guarantee that the heterogeneity of structural fiscal positions among regions in normal times is minimized.

  • Second, international experience shows that the occurrence of subnational fiscal crises cannot be ruled out, even in a setting in which national fiscal rules are fully credible and intergovernmental fiscal responsibilities are set at an optimal level. In the latter regard, the recent Spanish experience indicates that granting to regions additional instruments to prevent liquidity crises is warranted, so that pressure on the CG to financially support or bail out SNGs is reduced. In particular, the possibility of designing rainy day funds with regular contributions during periods of economic prosperity could be studied, along with the development of tools that guarantee the regular access of regions to financial markets even in periods of fiscal stress (Delgado-Téllez, González, and Pérez 2016).

Annex 14.1. Variables Used in the Empirical Analysis
Annex Table 14.1.1.Variables Used in the Empirical Analysis
VariableDescriptionSource
Fiscal Noncompliance Margin (official assessment)Difference between fiscal deficit targets and outcomes in percent of national GDP between 2003 and 2007, and in percent of regional GDP from 2008 to 2015Ministry of Finance
Fiscal Noncompliance Margin (homogeneous assessment)Difference between fiscal deficit targets (homogeneous assessement) and outcomes in percent of regional GDPAuthors’ calculation
Fiscal Deficit Targets (homogeneous assessment)Equal to Fiscal deficit targets (official assessment) × Nominal GDP (CG budget)/Regional GDP between 2003 and 2007, and to fiscal deficit target (official assessement) from 2008 to 2015Ministry of Finance (nominal and regional GDP); and authors’ calculation
Growth Forecast ErrorsReal GDP Growth Outturn – Real GDP ForecastNational Institute of Statistics (outturn), Ministry of Finance (forecast)
Regional-National Growth DifferentialRegional GDP Growth – National GDP GrowthNational Institute of Statistics
Regional-National Inflation DifferentialPercent Change in Regional CPI Growth – Percent Change in National CPINational Institute of Statistics
Fiscal Target AdjustmentDifference between fiscal defict target (homogeneous assessement) in the current and previous yearAuthors’ calculation
Execution Minus Budgetary Transfers (in regional GDP)Transfers from CG (outturns) – Transfers from CG (budget)Ministry of Finance and National Institute of Statistics
Regional Weight in National PopulationRatio of regional to national populationNational Institute of Statistics
Regional Weight in National GDPRatio of regional to national GDPNational Institute of Statistics
Regional Weight in National per Capita GDPRatio of regional to national per capita GDPNational Institute of Statistics
Tax AutonomyRatio of regional own revenues (regulatory power) to total regional revenuesMinistry of Finance; authors’ calculation
Social Spending Share in Regional Government SpendingRatio of regional spending in basic social services (health, education, and others) to total regional spendingInstituto Valenciano de Investigaciones Economicas and Ministry of Finance
Investment Share in Total Regional SpendingRatio of regional investment to total regional spendingMinistry of Finance
Vertical Fiscal Imbalances[1 – Regional Own Revenues/Regional Own Spending], where own regional revenue (spending) corresponds to a region’s total revenue (spending) minus transfers received by the CG and other public entitites (transfer paid to the CG and other public entitites)Authors’ calculation
Fiscal Rule IndexNumerical fiscal rule strength indexEuropean Commission
Fiscal Rule Index × Lagged Noncompliance MarginInteractions between the lag of noncompliance margin and the fiscal rule indexAuthors’ calculation
Region RatingsAverage rating numerical index, taking into account three rating agencies: Fitch, S&P, and Moody’sAuthors’ calculation using Fitch, S&P, and Moody’s databases
Implicit Interest RatesRegional interest payments in percent of end-of-year regional public debt stockMinistry of Finance
Ratio of Security to LoansRatio of total outstanding government securities issued by the regions to outstanding loans from commercial banksBank of Spain
National Election DummyDummy that equals 1 for the year of national parliament electionsWeb pages of the national and regional parliaments
Regional Election DummyDummy that equals 1 for the year of regional parliament electionsWeb pages of the national and regional parliaments
Party Congruence DummyDummy that equals 1 if regional and national government are led by the same party or party coalitionWeb pages of the national and regional parliaments
Proautonomy Party SharePercent of members of regional parliaments from regional pro-autonomy partiesWeb pages of the national and regional parliaments
Regions’ Seats in National ParliamentShare of members of the national parliament elected in each regionWeb pages of the national and regional parliaments
Source: Authors.Note: CG = central government; CPI = consumer price index.
Source: Authors.Note: CG = central government; CPI = consumer price index.
References

    ArgimónI. and P.Hernández de Cos.2012. “Fiscal Rules and Federalism as Determinants of Budget Performance: An Empirical Investigation for the Spanish Case.Public Finance Review40 (1): 3065.

    AuerbachA.1999. “On the Performance and Use of Government Revenue Forecasts.National Tax Journal52 (4): 76582.

    Autoridad Independiente de Responsabilidad Fiscal (AIReF).2016. “Informe sobre los presupuestos iniciales de las administraciones públicas para 2016,” April.

    BeetsmaR.B.BluhmM.Giuliodori and P.Wierts.2013. “From Budgetary Forecasts to Ex-Post Fiscal Data: Exploring the Evolution of Fiscal Forecast Errors in the EU.Contemporary Economic Policy31 (4): 795813.

    BeetsmaR.M.Giuliodori and P.Wierts.2009. “Planning to Cheat: EU Fiscal Policy in Real Time.Economic Policy24 (October): 753804.

    BoltonP. and G.Roland.1997. “The Breakup of Nations.Quarterly Journal of Economics11 (4): 105789.

    BordignonM.2006. ‘“Fiscal Decentralization: How to Harden the Budget Constraint.” In Fiscal Policy Surveillance in Europe edited by P.WiertsS.DerooseE.Flores and A.Turini.Basingstoke, U.K.: Palgrave Macmillan.

    BruckT. and A.Stephan.2006. “Do Eurozone Countries Cheat with Their Budget Deficit Forecasts?Kyklos59 (1): 315.

    CanutoO. and L.Liu eds. 2013. Until Debt Do Us Part: Subnational Debt Insolvency and Markets.Washington, DC: World Bank.

    CordesT.T.KindaP.Muthoora and A.Weber.2015. “Expenditure Rules: Effective Tools for Sound Fiscal Policy.Working Paper 15/2International Monetary FundWashington, DC.

    CoxG. and M.McCubbins.1986. “Electoral Politics as a Redistributive Game.Journal of Politics48 (2): 37089.

    CrivelliE. and K.Staal.2013. “Size, Spillovers, and Soft Budget Constraints.International Tax and Public Finance20 (2): 33856.

    Delgado-TéllezM.C. I.González and J. J.Pérez.2016. “Regional Government Access to Market Funding: International Experience and Recent Developments.Economic Bulletin[Bank of Spain] (February): 313.

    DixitA. K. and L.Londregan.1996. “The Determinant of Success of Special Interests in Redistributive Politics.Journal of Politics58 (4): 113255. European System of Accounts: ESA.2010. Euroopean Comission: Luxembourg. http://ec.europa.eu/eurostat/cache/metadata/Annexes/nasa_10_f_esms_an1.pdf.

    EyraudL. and L.Lusinyan2013. “Vertical Fiscal Imbalances and Fiscal Performance in Advanced Economies.Journal of Monetary Economics60: 57187.

    ForemnyD.2014. “Subnational Deficits in European Countries: The Impact of Fiscal Rules and Tax Autonomy.European Journal of Political Economy34 (June): 86110.

    FrankelJ. and J.Schreger.2013. “Over-Optimistic Official Forecasts and Fiscal Rules in the Eurozone.Review of World Economics149 (2): 24772.

    GrossmanP.1994. “A Political Economy of Intergovernmental Grants.Public Choice78 (3–4): 295303.

    HansenL.1982. “Large Sample Properties of Generalized Method of Moments Estimators.Econometrica50 (4): 102954.

    Hernández de CosP. and J.Pérez.2013. “Sub-National Public Debt in Spain: Political Economy Issues and the Role of Fiscal Rules and Decentralization.” In Fiscal Relations across Government Levels in Times of Crisis—Making Compatible Fiscal Decentralization and Budgetary Discipline edited by European Commission186216. European Commission Economics Papers 501, European CommissionBrussels.

    InmanR.2003. “Transfers and Bailouts: Enforcing Local Fiscal Discipline with Lessons from U.S. Federalism.” In Fiscal Decentralization and the Challenge of Hard Budget Constraints edited by J.RoddenG.Eskelund and J.Litvack3583. Cambridge, MA: MIT Press.

    International Monetary Fund (IMF). 2015. “Crisis Program Review.Policy PaperIMFWashington, DCNovember9.

    JonungL. and M.Larch.2006. “Improving Fiscal Policy in the EU: The Case for Independent Forecasts.Economic Policy21 (47): 491534.

    KotiaA. and VLledó. 2016. “Do Subnational Fiscal Rules Foster Fiscal Discipline? New Evidence from Europe.Working Paper 16/84IMFWashington, DC.

    Lago PeñasS.X.Fernández-Leiceaga and A.Vaquero.2016. “¿Por qué incumplen fiscalmente las CCAA?,FUNCAS Documento de Trabajo No. 784/2016MadridSpain.

    LealM. A. and J.López Laborda.2015. “Un estudio de los factores determinantes de las desviaciones presupuestarias de las comunidades autónomas en el periodo 2003–2012.Investigaciones Regionales31: 3558.

    LealT. and J. J.Pérez.2011. “Análisis de las desviaciones presupuestarias aplicado al caso del presupuesto del estado.Estudios de Economía Aplicada29: 909–32.

    LealT.M.Tujula and J.-P.Vidal.2008. “Fiscal Forecasting: Lessons from the Literature and Challenges.Fiscal Studies29 (3): 34786.

    Leite-MonteiroM. and M.Sato.2003. “Economic Integration and Fiscal Devolution.Journal of Public Economics87 (11): 250725.

    LledóV.2015. “Coordinating Fiscal Consolidation in Spain: Progress, Challenges, and Prospects.” In Spain Selected IssuesIMF Country Report 15/233IMFWashington, DC.

    Molina-ParraA. and D.Martínez-López.2015. “Do Federal Deficits Motivate Regional Fiscal (Im)balances? Evidence from the Spanish Case.GEN Working Paper A 2015-3University of Vigo, Santiago de CompostelaSpain.

    PérezJ. J. and R.Prieto.2015. “Risk Factors and the Maturity of Subnational Debt: An Empirical Investigation for the Case of Spain.Public Finance Review43 (6): 786815.

    Pérez GarcíaF.V.Cucarella and L.Hérnandez.2015. “Servicios Públicos, Diferencias Territoriales e Igualdad de Oportunidades.Fundación BBVABilbao.

    PortoA. and P.Sanguinetti.2001. “Political Determinants of Intergovernmental Grants: Evidence from Argentina.Economics and Politics13 (3): 23756.

    ReuterW.2015. “National Numerical Fiscal Rules: Not Complied with, but Still Effective?European Journal of Political Economy39 (September): 6781.

    RoddenJ.G.Eskelund and J.Litvack.2003. “Introduction and Overview.” In Fiscal Decentralization and the Challenge of Hard Budget Constraints edited by J.RoddenG.Eskelund and J.Litvack131. Cambridge, MA: MIT Press.

    SatoM.2007. “The Political Economy of Interregional Grants.” In Intergovernmental Fiscal Transfers: Principles and Practice edited by R.Boadway and A.Shah173201. Washington, DC: World Bank.

    SchaechterA.T.KindaN.Budina and A.Weber.2012. “Fiscal Rules in Response to the Crisis—Toward the ‘Next-Generation’ Rules. A New Dataset.Working Paper 12/187IMFWashington, DC.

    Simon-CosanoP.S.Lago-Peñas and A.Vaquero.2012. “On the Political Determinants of Intergovernmental Grants in Decentralized Countries: The Case of Spain.Working Paper 1230 International Center for Public PolicyGeorgia State UniversityAtlanta.

    StrauchR. M. and J.von Hagen.2009. “How Forms of Fiscal Governance Affect Fiscal Performance.” In Fiscal Governance in Europe edited by M.HallerbergS.Rolf and J.von Hagen.Cambridge, U.K.: Cambridge University Press.

    Ter-MinassianT.2015. “Promoting Responsible and Sustainable Fiscal Decentralization.” In Handbook of Multi-Level Finance edited by E.Ahmad and G.Brosio.Cheltenham, U.K.: Edward Elgar Publishing.

    VammalleC.D.Allain-Dupre and N.Gaillard.2012. “A Sub-Central Government Perspective on Fiscal Policy in a Tight Fiscal Environment.” In Institutional and Financial Relations across Levels of Government OECD Fiscal Federalism Studies edited by J.Kim and C.Vamalle1744. Paris: OECD/Korea Institute of Public Finance.

    VigneaultM.2007. “Grants and Soft Budget Constraints.” In Intergovernmental Fiscal Transfers: Principles and Practice edited by R.Boadway and A.Shah13371. Washington, DC: World Bank.

    von HagenJ.2005. “Political Economy of Fiscal Institutions.” In Oxford Handbook on Political Economy edited by D. A.Barry and R.Weingast46478. Oxford, U.K.: Oxford University Press.

    von HagenJ. and B.Eichengreen.1996. “Federalism, Fiscal Restraints, and European Monetary Union.American Economic Review86 (2): 13438.

    WildasinD.1997. “Externalities and Bailouts: Hard and Soft Budget Constraints in Intergovernmental Fiscal Relations.Policy Research Working Paper 1843World BankWashington, DC.

See Ter-Minassian (2015) for a review of this vast literature.

Attempts to address some of the flaws in the context of the European Union (EU), in particular strengthening fiscal rules without addressing others (for example, vertical fiscal imbalances), have been shown to be ineffective (Foremny 2014; Kotia and Lledó 2016).

See Argimón and Hernández de Cos (2012) for a review of this empirical literature.

Reuter (2015) shows that the introduction of numerical fiscal limits enforced through fiscal rules, even if not complied with, tilts fiscal policy outturns toward those numerical limits. So, in fact, compliance seems to matter less than whether the chosen numerical limit was set to an optimal or appropriate level.

In practice, fiscal target assessments usually occur at a time when factors underlying fiscal capacity, such as nominal GDP, are still only estimates.

Under an involuntary equilibrium, RGs must always be ex ante capable of complying with fiscal targets (that is, fiscal targets must be ex ante feasible). Ex ante infeasible fiscal targets could not be credibly enforced, fostering involuntary noncompliance.

A bailout is broadly defined to account for not only resources granted to SNGs in the event of a fiscal or financial crisis, such as emergency liquidity funds and outright debt restructuring, but also less extreme situations observed outside crisis. For instance, it may take the form of a change in the allocation of formula grants or simply unconditional gap-filling transfers. A bailout may include situations in which SNGs’ borrowing restrictions are lifted, allowing them to borrow to finance above-the-target fiscal deficit levels.

A critical assumption here is that the compliance assessment takes place before the bailout (that is, in the second stage). Bailouts that occur before the compliance assessment period (for example, gap-filling transfers) would help avoid or mitigate fiscal noncompliance. This requires corrective fiscal noncompliance measures or controlling the impact of alternative factors on uncorrected measures so as to take gap-filling transfers into account.

CG preference for bailing out politically aligned regions could also reflect electoral strategies to target safe electoral districts, that is, regions that had previously largely voted for and elected the CG party or governing coalition (Cox and McCubbins 1986). Such preferences may not necessarily prevail if the CG follows a swing strategy, whereby it will attempt to target regions that have previously voted for the CG party or governing coalition by narrow margins (Dixit and Londregan 1996). In some cases, such narrow margins may not have been sufficient for CG politically affiliated regional partners to win the election and form a government.

See Ter-Minassian (2015) for a review.

Large forecast errors, as discussed in the introduction, could also be the result of strategic considerations to ensure ex ante compliance with fiscal rules. In the context of the recent global financial crisis, they have also reflected larger-than-anticipated fiscal multipliers (IMF 2015).

This allows CGs to minimize the political costs of fiscal consolidations by preserving the provision of public goods and services under their mandate, while avoiding increasing the burden of their own taxes. CGs may also raise subnational fiscal targets to build buffers for possible noncompliance in different subsectors, RGs included.

Under the second BSL, fiscal targets were set in three stages. In the first stage, a report assessing the cyclical phase for the following three years was prepared. Taking into account the cycle, in a second stage, fiscal targets for the general government and subsectors (central, regional, and local governments as well as the Social Security System) taken together were set and submitted to Parliament. Once approved by Parliament and subject to the aggregate RG target, individual fiscal targets for each RG were set by means of bilateral negotiations between the Ministry of Finance and representatives of each RG on the Fiscal and Financial Policy Council.

Spain has 17 regions (comunidades autónomas). Nevertheless, two different center-periphery financial arrangements are in place. A majority of regions, 15, share the Common Regime of regional finances (comunidades autónomas de régimen común), with partial devolution of expenditure and revenues, while the remaining two (Navarre and Basque Country) enjoy a special status referred to as the Foral Regime of regional finances (régimen foral) under which they enjoy almost full spending and revenue autonomy. Within the latter two regions, though, the Basque Country is further decentralized, with revenue-raising responsibilities distributed to lower government levels (diputaciones forales) broadly resembling the provincial structure within the region. The latter region is therefore excluded from the subsequent econometric analysis because of the absence of comparable data.

Available at www.minhap.gob.es/esES/CDI/SeguimientoLeyEstabilidad/Paginas/InformesCompletosLEP.aspx. Two annual compliance assessments have been conducted since 2013. Noncompliance events are defined based on the second and final assessment.

The regional GDP series used is measured in market prices and in accordance with the new European System of National and Regional Accounts (ESA 2010).

The literature suggests that fiscal deficit at the CG level can encourage deficits at the RG level (see Molina-Parra and Martínez-López [2015] for the case of Spain) through the so-called copycat or yardstick effect. Nevertheless, this analysis did not find robust statistically significant evidence to support the hypothesis that fiscal compliance at the CG level influences fiscal compliance patterns at the subnational level. The results are excluded from the chapter for the sake of simplicity.

The key assumption here is that forecast errors are mostly driven by unanticipated changes in fundamentals and not by technical errors, weak or untimely data, or strategic motives (for example, overestimated nominal GDP growth forecasts to inflate revenue projections and make ex post excessive spending levels ex ante compatible with existing fiscal targets). Strategic motives and technical errors should play less of a role here to the extent that national growth forecasts are set by the center, where forecasting capacity and data quality are expected to be, on average, better than that of regions.

2010 and 2015 (widespread noncompliance and positive forecast error) were exceptions.

Adjustment effort could also be measured by the difference between fiscal deficit in year t and fiscal outturns in t−1. Unlike annual changes in fiscal targets, this measure is highly correlated with lagged fiscal compliance margins, and for this reason we have opted to exclude it from the baseline specification. Replacing it with our chosen adjustment effort proxy delivers qualitatively similar results at the expense of rendering lagged fiscal compliance margins statistically insignificant.

Regions have regulatory autonomy over personal income taxes (schedules, allowances, credits), wealth and estate taxes and property transfer taxes (schedules, deductions, credits), gambling (exemption, base, rate, credit), and vehicle registration (rates). Significant tax decentralization took place following the 1997, 2002, and 2009 reforms of the regional financing system.

Regions account for two-fifths of total general government spending on essential public services and more than 90 percent when it comes to health and education (Pérez García, Cucarella, and Hérnandez 2015), but about 5 percent with respect to social protection.

VFIs are defined as [1−Own Revenue/Own Spending ]. Own revenue (spending) corresponds to a region’s total revenue (spending) minus transfers received from RGs by the CG and other public entities (transfers paid by RGs to the CG and other public entities).

Although for VFI, noncompliance frequencies tended to increase up to the third quartile.

Although one cannot rule out the possibility of reverse causality, with fiscal noncompliance leading to poorer credit ratings, higher risk premiums, and costlier market financing.

As discussed in the section titled “Testable Hypotheses,” this may be the result of the CG following a “safe” electoral strategy. Simon-Cosano, Lago-Peñas, and Vaquero (2012) show that strategy to be preferred by national incumbents running in national elections, as reflected in the distribution of transfers to regions where the incumbent performs better.

    Other Resources Citing This Publication