Decentralization and the duration of fiscal consolidation: shifting the burden across layers of government

In the wake of the recent financial and economic crisis and during the ensuing recession, many OECD countries have opted to consolidate their general government budget balances. To achieve this goal, many have implemented severe austerity packages, with highly controversial results (Blyth 2013). General questions related to fiscal consolidation, including political economy considerations and macroeconomic conditions, have been studied extensively in the literature (Price 2010; Mauro 2011; Grüner 2013). Yet, despite the fact that many central governments recently have delegated substantial powers to sub-national entities (Rodríguez-Pose and Ezcurra 2011), the question as to whether fiscal decentralization strengthens governments’ ability to implement fiscal adjustment measures remains unanswered.¹

A key question concerning fiscal consolidation is its duration, a factor that is closely related to its effectiveness (see, e.g., Bi et al. 2013).² This appears to have been confirmed in the on-going debate as to which factors help sustain a consolidation effort. The point at stake seems to be the following: what is the optimal speed of fiscal consolidation and for how long should it be implemented? In other words, should governments concentrate their efforts at the beginning of the process and subsequently relax policy measures, or should they adopt the reverse approach? Blanchard and Leigh (2013) claim that these decisions depend on country-specific factors including, but not limited to, the level of public debt, economic growth, and interest rates. For example, austerity measures implemented in times of low or negative growth may damage the economy rather than help it recover (Ostry et al. 2016) and, in such cases, better strategies than upfront fiscal consolidation may be available.

The arguments for concentrating consolidation efforts at the beginning of the process have to be seen in the light of the default risks perceived by financial markets and of the costs of servicing public debt, since immediate fiscal adjustment measures might increase a government’s credibility and reassure market participants.³ This hypothesis is confirmed by Briotti (2004), who identifies persistent fiscal adjustment efforts as the best way to enhance credibility and ensure success as measured in terms of impact on deficits, debt, and growth performances. More recently, Barrios et al. (2010) have argued that sharp, sustained consolidations are more likely to be successful in countries with high starting debt levels and high interest rates (or low GDP growth potential), but that more gradual adjustments are preferable when these constraints are more relaxed.

However, despite the obvious importance of determining the optimum duration of a fiscal consolidation, no clear consensus has yet been reached. Thus, the hypothesis that decentralization can play a role in this process—in terms of how, to what extent, and with what results—is an appealing research question. A recent IMF policy paper (2014) recognizes the importance of intergovernmental fiscal coordination for the development of an effective consolidation plan. Indeed, institutional arrangements for coordinating financial decision-making between levels of government appear to be an increasingly important factor in the success of a government’s adjustment strategies.⁴ However, the fiscal objectives of most advanced economies tend to cover the general government sector only, and recognize no clearly identifiable contribution of sub-national levels of government to the targeted budget balance and debt position. Additionally, the distribution of the adjustment burden across the various layers of government is not pre-determined, and so it remains dependent on the strength of intergovernmental fiscal arrangements. An examination of the recent experiences of advanced countries seems to show that sub-national governments deliver proportionally larger shares of the overall reduction in the general government deficit than is delivered by the central government (Blöchliger 2013; Vammalle and Hulbert 2013). Our research explores this point in greater depth by analyzing the relationship between fiscal consolidation and decentralization. Should decentralization prove to be related to the duration of the former, this would have significant policy implications with different ramifications for federal and centralized countries.

From a theoretical viewpoint, fiscal decentralization may either favor or impede the governmental consolidation efforts. On the one hand, in a more decentralized system, the number of veto players increases, impeding the adoption of corrective packages and potentially deteriorating into a typical common-pool problem. Indeed, in times of crisis, a combination of deficit bias and coordination failures owing to fiscal decentralization may result in over-spending (and/or under-taxation) tendencies at the sub-national level as well as the adoption of conflicting fiscal stances at central and sub-central levels (see, e.g., Jonas 2012; Eyraud and Moreno Badia 2013; Foremny and von Hagen 2012).

Central governments, on the other hand, may consolidate their balances by simply cutting their inter-governmental transfers to sub-national units. In this way, they avoid having to face the direct political costs of such unpopular measures. The more expenditure is in the hands of sub-central tiers of government without corresponding revenue powers (as tends to be the case in most decentralized and advanced economies; see Blöchliger and Vammalle 2012), the more the burden can be shifted to sub-national levels. However, in these circumstances, the reactions of local governments to the negative revenue shock are crucial. If the reduction in transfers simply translates into an increase in sub-national deficits, no consolidation will have been achieved from a general government perspective. Today, various cooperative arrangements (including internal stability pacts and sub-national fiscal rules) have been put in place between central and sub-national governments (as exemplified by Austria, Germany and Italy), making sub-national constraints more binding and fiscal adjustment targets easier to achieve.

All of those possibilities depend heavily on the financing structure of the different levels of government: just how much autonomy do local governments really have over their own budgets? Foremny and von Hagen (2013) suggest that the actual degree of sub-central autonomy and reliance on transfers can have a critical impact on the relationship between decentralization and consolidation efforts. The authors report that sub-national governments in unitary countries experienced a significant increase in transfers from central government during the Great Recession, something that did not occur in the federal states. This suggests that the composition of transfers and changes in their configuration during periods of consolidation may depend on the degree of fiscal decentralization, and on the true assignment of powers to different tiers of government. In close relation to this, the presence of sub-national fiscal rules that limit certain budgetary items may also influence the duration and intensity of adjustment episodes. However, evidence of the disciplinary effects of sub-national rules (Nannicini et al. 2016; Escolano et al. 2012; Foremny 2014; Reuter 2015) has not yet been able to disentangle the discipline effect from the possibility of simultaneous bailouts from upper levels of government.

Our aim here, therefore, is to combine the literature on fiscal consolidation with that on fiscal federalism and decentralization. The specific question we address is whether more decentralized countries consolidate their budgets for longer periods than is the case with more centralized countries. We then examine the possible mechanisms behind our findings. More specifically, we study how central government transfers change during periods of consolidation as a means of determining whether such transfers are used by central governments to achieve their consolidation objectives. Finally, we analyze how countries with different systems of decentralization achieve their consolidation objectives in terms of the impact on public budget deficits and public debt, that is, we examine the relationship between fiscal decentralization and the success of the consolidation effort.⁵

Our main finding can be stated as follows: periods of fiscal consolidation are of longer duration in countries in which larger proportions of public spending are in the hands of sub-central tiers of government. This is especially true when such governments do not enjoy any real autonomy over their revenues. We also document a reduction in intergovernmental transfers during episodes of fiscal consolidation, suggesting that central governments shift the burden of consolidation onto lower levels of government whenever possible, at the expense of prolonging the consolidation process. Finally, we show that consolidation is, on average, less capable of improving the budget balance/GDP ratio in more decentralized countries, but at the same time is less damaging in terms of its impact on the public debt/GDP ratio in these countries. Our findings have interesting implications given the widespread concerns regarding the fiscal imbalances that currently characterize most developed economies.

The rest of the paper is organized as follows. Section 2 offers a brief literature review of fiscal consolidation, focusing on such aspects as its duration, determinants and success. Section 3 illustrates the empirical strategy we adopt to investigate our research question. Section 4 contains the results of the analysis, and Sect. 5 concludes.

Studies of fiscal consolidation have focused their attention on such issues as the factors that lead to the adoption of consolidation efforts, the determinants of their success, and the duration of the episodes. Existing empirical studies focus primarily on OECD countries, while interest in the subject has been revived by the recent economic and financial crisis and the calls heard for fiscal adjustments in most industrialized countries.

First, it should be stressed that no single definition of what constitutes a successful consolidation effort exists. For instance, Von Hagen and Strauch (2001) measure success in terms of the reduction in the budget deficit achieved at the end of the adjustment period. Alesina and Ardagna (2010) define a consolidation effort as having been successful when the cyclically adjusted primary balance (CAPB) as a share of GDP improves by 1.5% points or more. Barrios et al. (2010) distinguish between “cold shower” improvements of 1.5% points in the CAPB that are recorded in the space of one year and “gradual consolidations” that take place over 3 years if in each year the CAPB does not deteriorate by more than 0.5% of GDP.

An alternative definition considers a consolidation effort successful if the reduction in gross public debt achieved at the end of the episode has made the latter either sustainable or substantially smaller than it was at the outset (Heylen and Everaert 2000), although the debt/GDP ratio usually increases following a consolidation effort (Ardagna 2009) because of the adverse effects on the denominator of this ratio.⁶ Additionally, it might be claimed that the success of a fiscal adjustment program is related to the persistence of the consolidation effort: that is, for how long can (and should) a government consolidate its budgetary balances?

Changes in the expenditure/revenue mix would seem to play a key role in determining the outcome of a consolidation effort, although no consensus has emerged on the exact nature of those changes. On the one hand, spending-based adjustments are more likely to be successful, as they appear to be linked to longer lasting reductions in deficit/GDP ratios (e.g., Alesina and Perotti 1995; Afonso et al. 2006; Barrios et al. 2010; Alesina and Ardagna 2012). Devries et al. (2011) also claim that spending-based adjustments have been less contractionary in the past, but only because of accommodative monetary policy. At the same time, relying on higher taxes to reduce deficits may damage potential growth by discouraging labor market participation, and by lowering investment and firm profitability owing to the distortionary impact of taxes (especially those on income).

On the other hand, revenue-based consolidations may be more effective in terms of fiscal adjustments, particularly if they involve the revenues that are potentially less harmful for growth, that is, user fees, environmental taxes, property taxes and value-added taxes (Heylen and Everaert 2000; Tsibouris et al. 2006). In fact, revenue-based consolidations have been implemented in the past and several scholars recognize their effectiveness, especially when initial revenue-to-GDP ratios are relatively low. Moreover, it appears that the ex post composition of adjustments often turns out to be different from those originally planned, with expenditure cuts falling short of target and over-performing revenue changes (Tsibouris et al. 2006; Mauro 2011; Mauro and Villafuerte 2013).

In addition to changes in the expenditure/revenue mix, initial conditions also seem to matter in determining the success of fiscal consolidation efforts. In their seminal contribution, Von Hagen and Strauch (2001) investigate when fiscal adjustments are likely to be initiated, and under what circumstances consolidation efforts are likely to be successful, for a sample of European countries during the 1990s. The cyclical positions of the domestic and the international economy, the initial debt level, and the fiscal policy stance are all shown to be important determinants of the likelihood of a fiscal consolidation, as well as of its success. Barrios et al. (2010) also report that countries facing larger initial levels of government debt have a higher probability of pursuing successful fiscal consolidations. Cafiso and Cellini (2014) stress that although a certain debt/GDP ratio may affect the likelihood of consolidation, the opposite also is true, as consolidation affects debt dynamics. Their analysis of EU countries for the 1980–2009 period suggests that consolidation leads to lower debt/GDP values in the short-run, but not in the medium term.

On the other hand, Alesina and Ardagna (2012) find that initial conditions make no difference to the success or otherwise of a consolidation episode, while Devries et al. (2011) show that the role played by initial conditions is at best unclear. Molnar (2012) analyzes the economic environment, political settings and policy measures conducive to fiscal consolidation and debt stabilization and finds that the existence of fiscal rules and cooperation between different tiers of government play a critical role in favor of fiscal adjustment programs. The political framework also is relevant as newly elected governments seem more likely to initiate and sustain fiscal consolidation episodes, while non-centrist political parties are less likely to make efforts to stabilize debt than are those closer to the center of the political spectrum.

Another important dimension of a consolidation episode is its duration, an element that usually has been studied using survival analysis techniques. According to a European Commission study (2007), gradual consolidations tend to be more successful than quick, sharp adjustments, although the latter may be more effective in the case of high and rising debt levels. Below we summarize findings in relation to the duration of past fiscal adjustment efforts.

Von Hagen et al. (2002) were the first to concentrate on the duration of fiscal consolidation episodes and to make the length of the consolidation efforts endogenous in the empirical analysis. Their influential results highlight the importance of consolidation fatigue and of fiscal conditions such as public debt over GDP in determining the duration of fiscal consolidation. Illera and Mulas-Granados (2008) study the factors affecting the length of fiscal consolidation episodes, defined as the time spells between two fiscal expansions (which are in turn defined on the basis of the dynamics of the CAPB) in 15 European countries between 1960 and 2004. They find that the probability of a period of fiscal consolidation coming to an end, what the authors refer to as ‘failure’ (the term being taken from the standard survival analysis tools, where hazard functions are used to estimate the probability of a certain event, normally labelled as the ‘failure rate’), depends on such factors as the debt level, the magnitude of the adjustment, the relative contribution of spending cuts, and the degree of cabinet fragmentation.

It is worth noting that defining consolidation episodes solely on the basis of cyclically adjusted budget balance improvements may constitute too narrow an approach, since not all periods in which the balance did not improve should be considered failures (e.g., governments may simply not want to consolidate). In this respect, the data and definition provided by the IMF (Devries et al. 2011) seem to be more appropriate for investigating the duration of consolidation efforts. Here, fiscal consolidation episodes are classified using a narrative/historical approach based on the analysis of the policymakers’ intentions and actions as described in contemporary policy documents. Thus, the tax and spending measures taken in such periods are motivated primarily by the desire to reduce the budget deficit and not by a response to prospective economic conditions. We employ this type of data in our empirical analysis and so avoid the potential bias attributable to measurement errors that potentially correlate with economic developments and omit those consolidation efforts that are followed by adverse shocks that offset the discretionary measures. A further consequence of using this definition of consolidation is that its success is not ‘embedded’ in the definition, as the episodes in our dataset do not depend on numerical changes in the public deficit.

A more recent contribution to the study of the duration of consolidations is provided by Lodge and Rodriguez-Vives (2013), who estimate hazard functions for 20 advanced economies between 1970 and 2010. According to their analysis, the fiscal and macroeconomic conditions at the outset appear to affect the ability of governments to sustain lengthy consolidations. More precisely, high debt and deficits, heavy interest burdens, and high government bond yields all facilitate the initiation of consolidation (the so-called ‘push factors’); while, large private savings, strong external balances, competitiveness (measured by real exchange rate), and stable financial conditions facilitate its duration (the ‘pull factors’). Lodge and Rodriguez-Vives (2013) find that the composition of the fiscal adjustment (i.e., the split between expenditure and revenue measures) does not appear to be a significant determinant of the duration of consolidation.

Agnello et al. (2013), on the other hand, when using annual data for 17 industrial countries over the 1978–2009 period, find a difference between spending- and tax-driven consolidations, with the former episodes being shorter than the latter. Moreover, both types of consolidation are longer in non-European countries than they are in European countries, while the size of the consolidation program (in percentage of GDP) is not correlated significantly with its duration.⁷

As pointed out recently by the OECD (Vammalle and Hulbert 2013), successful national consolidation strategies usually benefit from involving sub-national governments. Moreover, to be successful, they also need to take into account the financial situation of these sub-national tiers so as to maintain the local authorities’ capacity to deliver major public services. Failure to do so can lead to breakdowns in coordination between different tiers of government, with adverse consequences for fiscal adjustment actions. Although it is acknowledged widely that sub-national governments are key players in fiscal policy-making, few studies of the durations of consolidations have paid close attention to their role or examined fiscal decentralization and intergovernmental fiscal relations.

Limited empirical evidence is available on this question. In the case of emerging market economies, Thornton and Adedeji (2010) find that sub-national governments have, in the past, contributed to successful general government fiscal adjustments by cutting their capital expenditures and raising their own tax revenues. In contrast, Baldacci et al. (2006) find no robust effects of fiscal decentralization (measured with simple dummies indicating the authority states and provinces have over fiscal policy) on the success of fiscal consolidation in a panel of 25 emerging market economies [note that they build on previous research on emerging economies by Adam and Bevan (2003) and Gupta et al. (2005)].

In the case of developed countries, Schaltegger and Feld (2009), in a case study of the Swiss cantons, find that fiscal centralization significantly reduces the probability of a successful fiscal consolidation, suggesting that competitive fiscal federalism may positively impact fiscal discipline. Darby et al.’s (2005) study of fiscal consolidations in OECD countries lends further support to the idea of the constructive participation of sub-national governments, with fiscal consolidations occurring at both central and sub-national levels. While other articles have investigated the role of fiscal federalism in the context of fiscal consolidation, they have not focused their attention on its relationship to the duration of the episode. It is this gap that we attempt to fill here with our original empirical research.

The need to adjust public finances and address government fiscal imbalances has generated a heated debate as to which fiscal tools (e.g., spending review measures, changes in tax systems) should be used and how they should be employed to restore sustainable fiscal positions. Research conducted to date has concerned itself primarily with analyzing the welfare and economic effects of the austerity measures implemented and with identifying the determinants of successful fiscal adjustment plans.

An important issue that has gone largely unexplored in these studies is the duration of these fiscal consolidation processes; yet, identifying the elements that might affect it constitutes an interesting research question. The IMF has suggested that fiscal institutions may be appropriate tools to sustain fiscal adjustment measures over time (Blanchard and Cottarelli 2010). Among such institutions, the organization of intergovernmental fiscal relations between the tiers of political authorities appears to be a natural candidate for influencing the length of a consolidation process. On the one hand, coordination failures and deficit bias problems may arise in the presence of multiple government tiers and agencies; however, on the other hand, more efficient and effective fiscal adjustments may result from the combined actions of several political actors in more decentralized systems operating under a cooperative institutional framework. Thus, we hypothesize here that the multi-layered fiscal structure, which characterizes most OECD countries, needs to be taken into account when assessing the durations of fiscal consolidations.

Our paper has analyzed the impact of fiscal decentralization on the durations of budgetary consolidation episodes in 17 OECD countries between 1978 and 2009. We find that: (1) consolidation episodes are longer in more decentralized countries, but only if sub-national governments have little real autonomy over their budgets; and (2) transfers from central government are cut during consolidation periods, and this effect is more pronounced if sub-national authorities have little legal power to influence central government decisions. Thus, the local governments that are at greatest risk from fiscal retrenchment episodes seem to be those that have a considerable number of spending tasks and responsibilities, but which are not accompanied by sufficient legal power.

Although we have not undertaken a formal welfare analysis, it seems that it may well be desirable for countries to implement decentralization reforms to mitigate the negative effects of fiscal consolidation. Gradual consolidation packages may allow the private sector to adjust more smoothly to government spending cuts without suffering any negative disruptions (see Cogan et al. 2013). Moreover, “a steady pace of adjustment” is, in general, less harmful for the recovery than a frontloading approach (Blanchard and Cottarelli 2010) and a sustained improvement in government fiscal balances is likely to crowd-in private domestic investment and net foreign assets in the long run (Anderson et al. 2014). In short, institutional settings that favor consolidation processes of longer duration are to be preferred. From this perspective, our results contribute to identifying a “new” advantage of fiscal decentralization. Hence, structural reforms in this direction coupled with fiscal consolidation episodes might offset the negative near-term implications of an austerity policy linked to fiscal adjustments.

The second piece of evidence revealed by our study concerns the implications our results have for the political economy. The fact that central governments reduce their transfers to lower tiers during fiscal adjustment processes may reveal a (possibly short-sighted) strategy of central governments to shift the adjustment burden onto lower government levels¹⁶ in order to appear virtuous in the eyes of both international markets and supranational institutions. This behavior runs contrary to the recommendations of the IMF (2013), which suggests that the best way to achieve credibility is to adopt medium-term fiscal plans with a visible anchor (e.g., either an average pace of adjustment or a fiscal target to be achieved within a certain period) combined with structural and institutional reforms, possibly involving different tiers of government.

Moreover, the central governments’ behavior of cutting their transfers to sub-national units during consolidation episodes may extend these episodes and lead to the loss of political consensus at the regional level, which, in turn, is likely to be reflected in the results obtained at national representative elections. To ensure the success of adjustment strategies, the IMF (2013), in fact, recommends a different approach based on the coordination of financial decision-making across different government tiers. The difficulties in implementing this coordination strategy may, however, be one of the reasons why more decentralized countries manage to achieve, on average, smaller improvements in the budget balance/GDP ratio than those recorded by their less decentralized counterparts (as we document in the last section of our analysis).

Furthermore, the distributional effects of such policies across government levels have to be given due consideration. All in all, given that budget consolidation measures normally are unpopular, the duration of a consolidation episode also is relevant in light of the political support governments require in order to implement such measures. In this regard, our findings suggest that the sub-national debt-to-GDP ratio increases when consolidation packages are implemented at the country level. This shift in the burden onto local public finances seems to be mitigated, however, by the real authority enjoyed by the sub-national government and, in particular, on its ability to participate and share in the central government’s policy-making and objectives, their implementation, and, ultimately, their success.