fiscal policy analysis in the euro area: expanding the toolkit

Please cite this article in press as: Paredes, J., et al. Fiscal policy analysis in the euro area: Expanding thetoolkit. Journal of Policy Modeling (2014), http://dx.doi.org/10.1016/j.jpolmod.2014.07.003

ARTICLE IN PRESS+ModelJPO-6146; No. of Pages 24

Journal of Policy Modeling xxx (2014) xxx–xxx

Available online at www.sciencedirect.com

ScienceDirect

Fiscal policy analysis in the euro area: Expanding thetoolkit�

Joan Paredes a,∗, Diego J. Pedregal b, Javier J. Pérez c

a European Central Bank, Germanyb U. Castilla-La Mancha, Spain

c Banco de Espana, Spain

Received 2 December 2013; received in revised form 22 May 2014; accepted 29 June 2014

Abstract

The absence of historical quarterly fiscal data has limited the analysis of the macroeconomic impact offiscal policies in the euro area, including the interactions of fiscal and monetary policies. To overcome thisgap, we construct a quite disaggregated euro area quarterly fiscal database for the period 1980Q1–2012Q4,based on a rich set of input fiscal data taken from national sources. We discuss how this dataset has allowedand can allow the profession to tackle new policy-relevant research topics. We also provide stylized facts onthe cyclical properties of main euro area fiscal aggregates, focusing on the recent economic crisis period.© 2014 Society for Policy Modeling. Published by Elsevier Inc. All rights reserved.

Keywords: Euro area; Fiscal database; Fiscal policies; Stylized facts; Mixed-frequencies’ models

JEL classification: C53; E6; H6

� The views expressed in this paper are those of the authors and not necessarily those of the European Central Bank or theBank of Spain. Initial versions of this paper circulated under the name: “A quarterly fiscal database for the euro area basedon intra-annual fiscal information”. We thank seminar participants at the European Central Bank, Jacopo Cimadomo, ToddClark, Günter Coenen, Giancarlo Corsetti, Francisco de Castro, Daniel Garrote, Domenico Giannone, Markus Kirchner,Michele Lenza, Albert Marcet, Henri Maurer, Agustín Maravall, Ad van Riet, Matthias Trabandt, and colleagues at theECB’s Fiscal Policies Division and Government Finance Statistics Unit, for useful comments and suggestions. We alsothank Lorenzo Forni, José Emilio Gumiel, Alexandru Isar, Sandro Momigliano and A. Jesús Sánchez for help with the data.Pedregal acknowledges financial support of the Spanish Education and Science Ministry under project SEJ2006-14732(ECON).

∗ Corresponding author. Tel.: +49 69 1344 5676; fax: +49 69 1344 7809.E-mail addresses: [email protected] (J. Paredes), [email protected] (D.J. Pedregal),

[email protected] (J.J. Pérez).

http://dx.doi.org/10.1016/j.jpolmod.2014.07.0030161-8938/© 2014 Society for Policy Modeling. Published by Elsevier Inc. All rights reserved.

dx.doi.org/10.1016/j.jpolmod.2014.07.003


http://www.sciencedirect.com/science/journal/01618938

mailto:[email protected]






2 J. Paredes et al. / Journal of Policy Modeling xxx (2014) xxx–xxx

1. Introduction

The European Central Bank (ECB) sets the monetary policy for the European economies thathave adopted the euro currency since January 1999. Therefore, macroeconomic analysis with euroarea time series became a common place over the past decade.1 The construction of historical data(long series) for the euro area has been part of the academic and ECB agenda over the first partof the 2000s, see Beyer, Doornok, and Hendry (2001), Anderson, Dungey, Osborn, and Vahid(2007), Fagan, Henry, and Mestre (2001, 2005). Even though fiscal policy remains a nationalresponsibility, interactions between monetary and fiscal policies are carefully monitored by themonetary authority, see for example, Duisenberg (2003), ECB (2008, 2009). In particular, theassessment of the impact of fiscal policies on euro area GDP and prices, and the constraints thatfiscal policies might impose on monetary policy over the medium term is a very relevant endeavor,increasingly so, in the light of the recent policy responses to the EU sovereign debt crisis. Euroarea governments introduced a number of discretionary fiscal policy packages: during 2008–2009fiscal stimuli and since the end of 2009 fiscal consolidation measures. Indeed, these issues haverecently attracted a great deal of attention.2 In addition, the analysis of spillover effects amongeconomic areas, in particular between the US and the euro area, or the euro area and the UK/therest of the EU are back to the forefront of the policy discussion.3

The appropriate assessment of the impact of fiscal policies at the euro area wide level and itsinterlinkages with other economic areas have been traditionally restricted by the shortcomingsof existing quarterly data for the relevant euro area fiscal variables. The whole fiscal surveillanceprocess at the European level is designed on the basis of annual data. The fact that budgetaryplans are prepared following an annual budgetary cycle and the discretionary nature of the setupchosen by many government taking measures for the entire year, have traditionally limited theinterest in high-frequency fiscal data.4

As recently claimed by Dilnot (2012) public policy analysis should not be undertaken lightlywithout thinking carefully and then finding out the numbers. Given the limitations and the scarcityof historical quarterly euro area fiscal data, we aim in this paper at reviewing existing, scatterednational data sources, and on that basis we construct a quarterly fiscal database for the euro areaaggregate5 for the period 1980Q1–2012Q4.6 The raw ingredients we use are closely linked tothe ones used by national statistical agencies to provide their best estimates (intra-annual fiscal

1 Examples are Batini, Callegari, and Melina (2012), Coenen, Straub, and Trabandt (2012), Cimadomo (2011a, 2011b),de Castro and Garrote (2012), Kollmann, Ratto, Roeger, and in’t Veld (2013), Burriel et al. (2010), Forni, Monteforte,and Sessa (2009), Ratto, Roeger, and in’t Veld (2009), Dreger and Marcellino (2007), Fagan et al. (2005), Favero andMarcellino (2005), Smets and Wouters (2003), Bruneau and de Bandt (2003), Aarle, Garretsen, and Gobbin (2003) orJacobs, Kuper, and Sterken (2003).

2 Just to quote a few examples, see Davig and Leeper (2011), Cogan, Cwik, Taylor, and Wieland (2009), Burriel et al.(2010), Cimadomo (2011a, 2011b), Cimadomo, Kirchner, and Hauptmeier (2010), or Coenen, Straub, and Trabandt(2013), Coenen et al. (2012). In the policy arena, ECB’s President introductory statement to the press conference typicallyincorporates an explicit reference to fiscal policies, see e.g. Draghi (2013).

3 See, for example, Auerbach and Gorodnichenko (2012).4 Nevertheless, a recent strand of the literature has shown that intra-annual fiscal data, when modeled appropriately,

contains valuable and useful information for forecasting annual aggregates (Pérez, 2007; Silvestrini, Salto, Moulin, &Veredas, 2008; Onorante, Pedregal, Pérez, & Signorini, 2010; Pedregal & Pérez, 2010; Asimakopoulos, Paredes, &Warmedinger, 2013; Leal, Pérez, Tujula, & Vidal, 2008).

5 The current euro area definition comprises the countries members of the euro area as of 31st December 2012.6 The database is updated once a year with the latest available data and can be requested at

euro area.fiscal [email protected].





J. Paredes et al. / Journal of Policy Modeling xxx (2014) xxx–xxx 3

data, mostly on a public accounts basis), and our method preserves full coherence with official,annual data. In order to make the database a usable input for applied empirical studies, includingthe estimation of macroeconomic models7 we provide a quite disaggregated set of fiscal variablesfor the General Government sector.8

In fact, the potential for policy applications of our database has been proved in a number ofrecent papers that could not have been completed as they stand had our set of data not beendeveloped (see, among others, Burriel et al. (2010), Batini et al. (2012), Coenen et al. (2012,2013), Cimadomo (2011a,b), de Castro and Garrote (2012), Kollmann et al. (2013), EuropeanCommission (2012)).9

In addition, taking advantage of the dataset, and moving one step forward, we also offerin our study a description of fiscal policy developments in the euro area including the crisisyears (2009Q1–2012Q4).10 In the historical period considered, that covers a number of economicdownturns such as the 1980s, the mid-1990s, the 2000s and the most recent crisis, only thelatest, also called the Great Recession, caused total euro area government revenues to enter intonegative territory in nominal terms. The fiscal adjustment process that took place was mainly onthe revenue side of the budget: indeed, the public deficit reduction between 2010Q4 and 2012Q4was principally due to total revenues, which increased by 2.9 percentage points of 2012’s GDPover that period, and compensated the 0.6 percentage points of 2012’s GDP increase in totalexpenditure. Within total expenditure, though, a substantial reduction in public investment tookplace (−0.4 percentage points of 2012’s GDP), that partly compensated the boost in currentexpenditure, which can be attributed to the effect of automatic stabilizers. Thus, from the point ofview of the composition of the fiscal adjustment, the features described in this paragraph wouldnot be, inline with the available empirical and theoretical literature, the best in terms of impacton economic growth and conductive to successful consolidation (see e.g. Alesina and Ardagna(2010), Bi, Leeper, and Leith (2013)).

Finally, we also provide detailed stylized facts on fiscal policies for the euro area on the basisof the dataset, for the whole sample (1980Q1–2012Q4) and for two relevant subsamples, namely,the pre-crisis period (i.e. the whole sample excluding 2008Q1–2012Q4) and the euro area period(since 1992Q1, i.e. including the run-up to the euro). First, as expected, we find a strong and pro-cyclical behavior of total government revenue in the euro area, which follows the business cyclebehavior in upturns and downturns, reflecting the operation of automatic stabilizers. When the2008Q1–2012Q4 period is included the synchronicity of public revenues and real GDP increasescompared to the pre-2008 sample. Second, and more interesting, are the results on the cyclicalproperties of government spending. When comparing the three analyzed samples, it seems thatthe crisis and subsequent fiscal consolidation period (2008–2012) reduced the pro-cyclical biasof public spending. On the one hand, some current spending items did not fall sharply during theGreat Recession (2008–2009). On the other hand, during the most recent period (2010–2012),those items did not increase much, or even decreased due to the subsequent fiscal adjustment

7 Like ECB’s AWM and NAWM, see Fagan et al. (2001, 2005), and Coenen, Christoffel, and Warne (2008), respectively.8 In National Accounts’ (NA) terms, i.e. according to ESA95 (European System of National Accounts) definitions.9 In addition, since the September 2010 edition of ECB’s euro area AWM database, its initial fiscal block has

been substituted by the early vintages of the historical dataset presented in the current paper. See “The AWMdatabase”, September 2010, available at the official AWM site with the Euro Area Business Cycle Network(http://www.eabcn.org/data/awm/index.htm).10 The studies mentioned in the previous paragraph used the beta version of the dataset that only covered the pre-crisis

period.


http://www.eabcn.org/data/awm/index.htm




implemented by euro area economies. The latter result is particularly relevant in the case ofpublic investment.

The rest of the paper is organized as follows. In Section 2 we examine issues related to theavailability of fiscal data for the euro area and the main features of the data used in our study (thedetails of which are further developed in Appendices A and B).11 In that section we also offersome descriptive evidence on the evolution of fiscal variables in the euro zone over the 1980–2012period. In Section 3 we discuss the relevance of our dataset for the study of fiscal policy issues,first by showing how we deal with potential endogeneity problems, and second by reviewing therecent applied literature. In Section 4, in turn, we provide stylized facts on the cyclical propertiesof the main fiscal aggregates. Finally, Section 5 concludes.

2. Euro area fiscal data: fiscal variables over the economic crisis and the fiscalconsolidation episode

2.1. General issues

The euro area is an aggregation of member states’ country-specific time series. There is nothinglike a “euro area fiscal policy”. Nevertheless, as claimed before, the monetary policy of theECB is conducted taking into account euro area fiscal aggregates as if they were representing asingle entity/country. Thus, the ECB has devoted a great deal of effort in building up consistentdatabases of country-specific data and in the development of aggregation tools to assemble euroarea aggregates for the different macroeconomic variables. In the fiscal domain, Eurostat andthe ECB provide annual series for euro area fiscal aggregates that dates back to the 1995s. Inaddition, Eurostat, on the basis of data provided by EU National Statistical Institutes, providesdisaggregated quarterly non-seasonally adjusted12 government data for the euro area for the periodstarting in 1998Q1. The compilation practices follow the guidelines of the manual on quarterlynon-financial accounts for general government (see European Commission (2006)). Using thelatter accounting approach to extend back in the past existing euro area fiscal time series is not afeasible endeavor, though, given the limited information available.

The main sources of intra-annual fiscal data in the euro area covering long periods of time arenational sources. Most countries publish on a monthly and/or quarterly basis, for example, centralgovernment accounts. In Federal or quasi-federal countries, like Germany and Spain, regionaland local government finances at the quarterly/monthly frequency are also covered, even thoughthe details tend to be lower than the corresponding central government counterparts. With theexception of the United Kingdom (a country within the EU but outside the euro area), and to alesser extent France, it is not possible to find official time series with a wide time and institutionalcoverage of the General Government sector in national accounts – see Onorante et al. (2010), orPedregal and Pérez (2010) for some additional discussion on these issues.

2.2. Main features of the data

A number of issues have to be dealt with carefully and explicitly for any set of non-officialdata (i.e. data not produced regularly by a National Statistical Institute or with an official status

11 An additional document with further information is available from the authors upon request.12 Seasonal adjusted data started to be published recently by Eurostat only for total revenues and total expenditures.





by an international organization) to be trustworthy, and thus usable by a wide audience. These arethe main characteristics of the database used in the paper (see Appendices A and B for further,technical information):

Input fiscal data. Our database is built up using only intra-annual fiscal information, i.e.we do not use general quarterly macroeconomic variables – like GDP, private consumption ortotal economy employment – in the interpolation process. This is a quite relevant issue becausealthough government revenues and expenditures (e.g. unemployment benefits) may be endoge-nous to GDP or any other tax base proxy (e.g. private consumption for VAT collection) therelationship between these variables is at most indirect and extremely difficult to estimate. Thedecoupling of tax collection from the evolution of macroeconomic tax bases (revenue wind-falls/shortfalls) is by now a proved stylized fact. In this respect, the direct use of intra-annual fiscaldata, taken from public accounts’ sources, for interpolation purposes, might be instrumental toavoiding the potential problem of modeling an indirect relationship which, in addition, might betime-varying.13

Compilation approach. We choose in this paper an econometric approach rather than anaccounting approach.14 Nevertheless, we tried to follow to the extent possible the principlesoutlined in the manual on quarterly non-financial accounts for general government: use of directinformation from basic sources (public accounts’ data), computation of “best estimates”, and con-sistency of quarterly and annual data. In this respect, we chose intra-annual data from the publicaccounts of the individual countries, along the lines of the statement of the manual that quarterlydata shall be based on direct information available from basic sources, such as for example publicaccounts or administrative sources. A description of the main fiscal indicators used in the studyis described briefly in Table 1 (see also Appendix A).

Aggregation of euro area data. The approach followed in our paper is an indicator-based one.This means that we do not aggregate data of the individual euro area member states as such.Instead, we use aggregated annual data as provided by the European Commission (Eurostat) and(when available) quarterly euro area data from the same source, as anchors for the interpolationprocedure, while at the same time we set up statistical models that incorporate ingredients thatclosely resemble those used to compile available quarterly government finance statistics databy Eurostat, for the biggest euro area economies, namely Germany, France, Italy, Spain and theNetherlands. We do so for two main reasons. Firstly, to maximize data availability, and in particular,the length of the available series, an aggregation-based approach would have blocked many timeseries, and seriously limited the length of the feasible ones. In this respect it is worth mentioningthat all the ingredients of the dataset are publicly available, i.e. we made no use of restricted or

13 In a related fashion, the use of quarterly fiscal indicators in the interpolation, should be of use in capturing accuratelythe quarter(s) in which, for example, a change in a given tax rate took place, an issue that is crucial to assess the impactof fiscal policies on GDP and other macroeconomic aggregates, not least because of the existence of foresight effects.14 The discussion in European Commission (2002a, 2002b, 2006) shows that there is some room for econometric

estimation of intra annual fiscal variables. This is the case for two main reasons, highlighted in the previous references.Firstly, ESA95 does not consider the quarterly aspects of taxes and social payments with sufficient precision to ensureclarity of interpretation in all situations; this is because, when discussing non-financial accounts, the ESA95 guidingdocuments occasionally take a perspective which assumes an annual reference period is in mind, thus remaining silent onwhich quarter within a particular annual reference period is involved. Secondly, it is also the case that many accounting orlegal events are annual events by definition (e.g. a tax levied in a complete year); this fact does not present a problem forthe statistician compiling annual data (there is no need to establish the amount and time of recording to a particular annualreference period), but do pose problems for the compiler of quarterly data, that needs to attribute revenue and expenditurenot merely to a reference year but also to quarters within that year.


Please cite

this article

in press

as: Paredes,

J., et

al. Fiscal

policy analysis

in the

euro area:

Expanding

thetoolkit.

Journal of

Policy M

odeling (2014),

http://dx.doi.org/10.1016/j.jpolmod.2014.07.003

AR

TIC

LE

IN P

RE

SS

+Model

JPO-6146;

N

o. of

Pages 24

6

J. Paredes

et al.

/ Journal

of Policy

Modeling

xxx (2014)

xxx–xxxTable 1Overview of main sources of national government’s monthly, quarterly and annual data, and of aggregate euro area annual and quarterly series.

Germany France Italy Spain The Netherlands Euro area aggregates

Total revenue TOR Federal Ministry ofFinance, Eurostat, Bankof InternationalSettlements (BIS)

Ministry of Finance,Eurostat

Banca d’Italia IGAE (State Comptroller),National Statistical Institute(INE)

Ministry of Finance Eurostat

Direct taxes DTX Federal Ministry ofFinance, Eurostat

Bank of InternationalSettlements (BIS),Eurostat

Eurostat National Statistical Institute(INE), Eurostat

Eurostat Eurostat

Corporate taxes DTE OECD OECD OECD OECD OECD EurostatSocial securitycontributions

SCT Bank of InternationalSettlements (BIS)


– National Statistical Institute(INE), Eurostat, SocialSecurity System

– Eurostat

Indirect taxes ITX Bank of InternationalSettlements (BIS),Eurostat


Eurostat Eurostat Bank of InternationalSettlements (BIS),Eurostat

Eurostat

Total expenditure TOE Federal Ministry ofFinance, Eurostat, Bankof InternationalSettlements (BIS)

Ministry of Finance,Eurostat

Banca d’Italia IGAE (State Comptroller),National Statistical Institute(INE)

Ministry of Finance Eurostat

Interest payments INP Federal Ministry ofFinance

Bank of InternationalSettlements (BIS)

Giordano et al.(2007), Eurostat

IGAE (State Comptroller),National Statistical Institute(INE)

– Eurostat

Governmentconsumption

GCN Eurostat, Bank ofInternational Settlements(BIS), Eurostat

Eurostat Eurostat Eurostat Eurostat Eurostat

Real GCR Eurostat, Bank ofInternational Settlements(BIS), Eurostat

Eurostat Eurostat, Bankof InternationalSettlements(BIS), Eurostat

Eurostat, Bank ofInternational Settlements(BIS), Eurostat

Eurostat Eurostat

Governmentemployment

LGN Eurostat Eurostat Eurostat Eurostat Eurostat OECD, Eurostat, Pérezand Sánchez (2011)

Compensation ofemployees

COE Federal Ministry ofFinance




– Eurostat

Governmentinvestment

GIN Eurostat, Bank ofInternational Settlements(BIS), Eurostat




Eurostat, Bank ofInternational Settlements(BIS), CBS

Eurostat





Table 2Coverage and structure of the quarterly fiscal dataset.

Deficit (DEF) = TOR − TOE

Total revenues (TOR) Total expenditure (TOE)

Direct taxes (DTX) Social payments (THN)-Paid by enterprises (DTE) -of which unemployment benefits (UNB)-Paid by households (DTH = DTX − DTE) Government consumption (GCN)

Social security contributions (SCT) -Compensation of employees (COE)-of which paid by employers (SCR) -Non-wage consumption expenditure (OGCN = GCN − COE)-of which paid by employees (SCE) Subsidies (SIN)

Indirect taxes (TIN) Government investment (GIN)Other revenues (OTOR = TOR − DTX − SCT − TIN) Interest payments (INP)

Other expenditure (OTOE = TOE − GCN − SIN − GIN − INP)

Government employment (LGN)Real government consumption (GCR)

private information. The second reason is to avoid the controversial issues of weighting schemes,as discussed in Beyer et al. (2001), Bosker (2006), Brüggemann and Lütkepohl (2006) or Andersonet al. (2007).

Coverage. The database covers the period 1980Q1–2012Q46 above, thus it includes the mostrecent crisis period. It encompasses the main components of the revenue and expenditure sides ofthe General Government sector (see Table 2) in NA terms. The net lending of the government, a keypolicy variable can be computed as the difference between total revenues and total expenditures.In addition, we also provide general government debt.

Definitions. We provide seasonally adjusted series, which are consistently and jointly estimatedwithin our models. The issue of seasonal adjustment of quarterly fiscal variables in Europe is animportant one, as signaled in European Commission (2007). Currently, available disaggregatedquarterly government finance official figures are presented mainly in non-seasonally adjustedterms, given the short time span available (the starting period is 1999Q1), two features thatmake difficult the economic analysis with those figures. Indeed, adjusting in a robust way forseasonality such short time series is a difficult endeavor. In this sense, given that we use a broadset of information and model explicitly seasonality for the whole set of series included in ourmodels, for the period 1980Q1–2012Q4, we are in a position to provide, in particular, seasonallyadjusted series computed in a robust way for the period for which the official statistics are available(1998Q1 onwards).

2.3. Some descriptive evidence

Figs. 1–3 present year-on-year growth rates of the main general government variablesfor the euro area aggregate (seasonally adjusted). Fig. 1 shows total government rev-enue and its components, Fig. 2 total government expenditure and its components, andFig. 3 zooms in the decomposition of government consumption into wage and non-wageexpenditures.

As regards the information displayed in Fig. 1, it is apparent that the most recent crisis hadthe largest negative impact on total government revenues in the analyzed sample. Indeed, in a





Fig. 1. The evolution of total public revenue and its components in the euro area, 1980Q1–2012Q4. Year-on-year growthrates of seasonally adjusted figures in nominal terms.





Fig. 2. The evolution of total public expenditure and its components in the euro area, 1980Q1–2012Q4. Year-on-yeargrowth rates of seasonally adjusted figures in nominal terms.





Fig. 3. The evolution of government consumption and its components in the euro area, 1980Q1–2012Q4. Year-on-yeargrowth rates of seasonally adjusted figures in nominal terms.

historical period, which covers a number of economic downturns such as the 1980s, the mid-1990s, the 2000s and the most recent crisis, only the latest, also called Great Recession, causedtotal euro area government revenues to enter into negative territory in nominal terms. In fact, totalrevenues contracted for five consecutive quarters, namely since 2008Q4 till 2009Q4, presentingand average drop of 5% per quarter in year-on-year terms. The main components of public revenuespresented a similar profile over the crisis, but with significantly different amplitudes. Direct andindirect taxes presented negative rates of growth for six consecutive quarters, while Social Securitycontributions displayed a more moderate drop, consistent with the more stable tax bases that arethe source of such government incomes. In any case, the fall in social contributions is the biggestin the sample after the one observed at the end of the 1990s that was due to significant cutsin employees’ contributory rates in France. The significantly higher volatility of direct taxes ascompared to other components is completely driven by the behavior of corporate tax receipts (seepanel with total direct tax collection, corporate tax revenues and personal income tax revenues).Indeed, the relative standard deviation of direct taxes with respect to total revenues equals 1.5,while that for indirect taxes and social contributions is equal to 1.0. The most volatile componentof revenue, in any case, is the residual aggregate “Other government revenue”, given that iscomprises an aggregation of items much smaller in magnitude and more subject to discretionary





impact, namely capital revenues, other current revenues (including interest receivable) and sales(non-market production).

The crisis did not have, in any case, a significant impact on the percentage structure of govern-ment revenues in the euro area. Direct taxes, social contributions and indirect taxes representedin 2012 a share equal to 27%, 34% and 28% of total revenues, respectively, compared to 28%,33% and 29% in 2007. Nevertheless, within direct tax collection the share of corporate taxes fellby 6 percentage points.

As regards the aftermath of the 2008–2009 crisis, that coincided with the period of fiscaladjustment (generalized tax increases) that started in mid-2010, it is worth noticing that totalrevenues at the euro area level increased by 6.7% between 2010 and 2012 (compared to +0.7%of nominal GDP), in such a way that in cumulative terms in 2012Q4 an extra amount of D 276 bnwas collected by euro area governments when compared with the year 2010, i.e. 2.9 percentagepoints of 2012’s GDP.

In Fig. 2, in turn, we show total expenditure and its components. Government consumption andcash transfers to households represent the bulk of total expenditure, with shares (in 2012) of 43%and 35% respectively; the ratio of the standard deviation of consumption expenditure and transferscompared to the standard deviation of total spending stands at 1.2 and 1.0, respectively. The smallercomponents, in turn, present much higher relative volatility with respect to the aggregate, of 4.3for government investment (5% weight), 3.3 for interest payments (6% weight), and, particularly,of 9 for “Other expenditures” (computed as a residual and amounting to some 8% of aggregatespending). On a related fashion, expenditure in subsidies is a small item amounting to some 3%of the total, and with a relative standard deviation that doubles that of the aggregate. We also showin the figure unemployment benefits, a subcomponent amounting to some 8% of social transfers,and some 4 times more volatile than this latter aggregate.

Within government consumption, as shown in Fig. 3, non-wage consumption expenditure ismore volatile than wage expenditure (compensation), 1.8 and 0.9 in terms of relative standarddeviations to government consumption respectively, while both amount to a similar share of totalconsumption (some 50%).

From an aggregate perspective, Figs. 2 and 3 provide an overview of the composition ofthe fiscal consolidation effort implemented by individual euro area countries. The expenditureadjustment process started with intensity approximately in 2010Q3, dominated by the decelerationin government consumption and significant cuts in government investment. As regards the latter,government investment expenditure presented the most prolonged and intense period of declinein the sample (including the 1980s), with negative average year-on-year nominal growth rates ineach single quarter of −7% in 2010Q1–2012Q4. As regards the former, the fall in expenditure incompensation of government employees was the main driver in the nominal deceleration – andultimately reduction – of consumption expenditure, in particular its real component, includingpublic employment cutbacks.

Overall, nevertheless, between 2010 and 2012, total euro area government expendituresincreased by D 56 bn (+1.2% vs. the +0.7% of nominal GDP), due to the fact that the upsurgein social transfers (+D 64 bn) and interest payments (+D 35 bn) was only partly compensatedby the nominal reductions in government investment (−D 34 bn), the wage bill (−D 5.9 bn) andother expenses (−D 3 bn). Overall, thus, 79% of the expenditure adjustment was due to cuts ingovernment investment, a composition that has been typically advocated as neither being theless harmful for economic growth nor the most conductive to a successful fiscal consolidationprocess (within a huge literature, see e.g. Perotti (1996), Alesina and Ardagna (2010), Bi et al.(2013)).





3. The relevance of the dataset for the study of fiscal policy issues in the euro area

There are a number of areas of research and applied policy analysis in the field of fiscal policiesin the euro area which can benefit from the detailed dataset presented in the current paper. In thissection, first we outline some claimed advantages of the proposal we put forward in this paper and,second, we provide some examples of applications prepared with our dataset by other authors.

3.1. Fiscal revenues and theoretical macroeconomic tax bases

In any standard macroeconomic model in which taxes are included, tax collection tends to belinearly related, via a tax rate, with a theoretical tax base. Let us take the example of consumptiontax collection (Tc): Tc = τcC, where τc is the implicit tax rate and C denotes private consumption.The latter is typically measured by the relevant National Accounts’ (NA) aggregate. For modelconsistency, this approach is a valid one. But from a practical point of view, it presents a number ofproblems. First, the definitions of taxes in OECD economies tend to encompass a number of taxablecategories that are not always properly captured by NA main aggregates. In the previous example,following Eurostat (EU’s Statistical Agency), indirect taxes, defined as taxes on production andimports, comprise mainly tax categories such as VAT (the major item), property taxes, exciseduties (such as taxes on gasoline and other fuels, and taxes on tobacco and alcohol) and taxes andduties on imports excluding VAT, and thus the relevant tax base should comprise in addition to NAprivate consumption, other variables like consumption of fuel, tobacco and alcohol, residentialinvestment, house purchases, government intermediate consumption, certain imports, or exportsof services like tourism. A second issue is related to the fact that tax systems are complex functionsof the tax bases that determine the revenue responsiveness properties of different taxes to the stateof the economy (see, for example, Creedy and Gemmell (2002, 2007)).

These features, among others, make the relationship between tax collection and spend-ing of endogenous items (like unemployment benefits) and their theoretical macroeconomictax/spending bases a non-exact, indirect one. As an example, in Fig. 4(Panel A) we show theannual growth rates of indirect tax collection and private consumption, both deflated by the pri-vate consumption deflator. Even though average growth over the three decades displayed in thechart is broadly the same among the two series, the volatility of the growth rate of indirect taxes wasalmost 2 times higher than that of private consumption. Just focusing on the last decade, indirecttax collection was above this proxy macroeconomic base in the “good times” period 2002–2007,and quite below it in “bad times” (2000–2001, 2008–2009). Thus, in good times a typical indirecttax revenue equation would present positive residuals, while it would display negative residu-als in recessions. This phenomenon is typically referred to as revenue windfalls/shortfalls. Morespecifically, the term revenue windfalls/shortfalls is usually used in the relevant literature15 todescribe government revenues which fall short of (are in excess of) what would be expected inview of the impact of legislation changes and the actual or projected development of standard keymacroeconomic aggregates.

The database presented in our paper has been constructed by using only direct information onintra-annual fiscal developments. To give a flavor of the relevance of our approach compared to

15 See, among others, Barrios and Rizza (2010), Morris et al. (2009) and Morris and Schuknecht (2007). See also Penner(2008). A related literature is the one on “rainy day” funds. Budget stabilization or “rainy day” funds is a practice followedby many US states that consist of setting aside excess revenue for use in times of unexpected revenue shortfall. In fiscalyear 2008, forty-seven states and the District of Columbia maintained rainy day funds.





Fig. 4. Real private consumption and real indirect tax collection. Panel A: Annual data (year-on-year growth rates),1981–2012. Panel B: Cross-correlation function of real private consumption and real indirect tax collection (quarter-on-quarter growth rates), by means of two interpolation alternatives. Sample 1980Q2–2012Q4.





the alternative of interpolating annual fiscal accounts with quarterly macro indicators (as in thecases of the editions of ECB’s AWM database published prior to 2009 or the companion datasetin Forni et al. (2009)), we present the following exercise. First, we compute quarterly series of(seasonally adjusted, real) indirect tax collection using the same econometric methodology asin our database but using only quarterly private consumption as the indicator to distribute theannual figures among quarters. Second, we compute the cross-correlation functions (CCFs) of thequarter-on-quarter growth rates of the indirect tax series obtained with the two alternatives (fiscalindicators versus private consumption) and real private consumption for 4 leads and lags. Thetwo CCFs are plotted in Fig. 4(Panel B). It is clear that the second alternative provides a highercontemporaneous correlation of consumption tax collection with NA private consumption.

As a conclusion, one may claim that approaches that use macroeconomic indicators to inter-polate fiscal series may inflate the underlying “true” relationship between the so-constructedquarterly fiscal series and the headline macro figures. Maybe this cost is worth assuming in cer-tain studies for the sake of model consistency, but it would certainly harm estimated relationshipsin empirical studies.

3.2. The impact of fiscal policies in the euro area economy

Given the potential importance of the spillover effects of fiscal policies in a highly integratedregion such as the euro area, the results available for some specific countries16 do not necessarilyprovide a good guidance for analyzing the macroeconomic impact of fiscal shocks in the euro areaas a whole.17 While several studies have focused on the United States, results for the euro areahave been scarcer, primarily because of lack of data availability. Some recent empirical studiesstress this data availability problem, and use the historical dataset described and presented in thecurrent paper to extend the set of available facts on the effects of government shocks on euroarea GDP and inflation: Cimadomo et al. (2010) and Burriel et al. (2010), Batini et al. (2012) orEuropean Commission (2012).

From a more structural point of view, some additional recent studies, namely Coenen et al.(2012, 2013), make use of the database presented in this paper to estimate DSGE models forthe euro area. In particular, in the second paper the authors conduct a quantitative evaluation ofdiscretionary fiscal policy on euro area economic activity during the Great Recession, and to thisend, they use a DSGE model characterized by a rich specification of the fiscal sector and estimateit utilizing a large set of euro area fiscal data.18

16 For euro area country studies see, among others, Heppke-Falk, Tenhofen, and Wolff (2006) and Baum and Koester(2011) for Germany, de Castro (2006) and de Castro and Hernández de Cos (2008) for Spain, Giordano, Momigliano,and Neri (2007) for Italy, Marcellino (2006) for the four largest countries of the euro area or Afonso and Sousa (2009a,2009b) for Germany, Italy and Portugal, and Bénassy-Quéré and Cimadomo (2012) and Beetsma and Giuliodori (2011)for a group of EU countries. On different grounds, Jacobs, Kuper, and Verlinden (2007) incorporate a fiscal closure rulein a VAR for the euro area.17 For empirical studies that include EU countries and focus on cross-country spillovers see Canzoneri, Cumby, and

Diba (2006), Beetsma, Giuliodori, and Klaasen (2006), Cwik and Wieland (2010), and Corsetti, Meier, and Müller (2010)for theoretical considerations. See also Pappa (2009) that compares the transmission of government spending shocks inCanada, Japan, the UK, the US and the euro area. See also Canova and Pappa (2011).18 In order to address the potential problem of mismeasurement associated with the use of interpolated data, the authors

allow for errors in the measurement of the fiscal variables. In particular, for all fiscal data iid measurement errors with avariance of 0.5% are assumed.





4. The cyclical properties of the main fiscal aggregates in the euro area

We provide in this section the cyclical properties of the main fiscal aggregates in the euroarea. We have chosen this topic because of its policy relevance and because there are relativelyfew empirical studies on the pro- versus counter-cyclicality of fiscal policies for the euro area.Previous studies have typically focused on a limited number of euro area or euro area countries’fiscal aggregates and used annual data (Fatas & Mihov, 2009; Lamo, Pérez, & Schuknecht, 2013;Marcellino, 2006). The studies find mostly pro-cyclical fiscal behavior. The most comprehensivestudy of public expenditure cyclicality in the OECD is Lane (2003), who estimates the elasticityof government expenditure and its components with respect to output for a number of countriesfor the period 1960–1998. Lane finds that government consumption in most countries behavespro-cyclically, mainly due to the behavior of wages.

In Table 3, we report dynamic cross-correlation functions. We look at the unconditional corre-lations between detrended series at the standard business cycle frequencies. Following standardpractice we measure the co-movement between two series using the cross-correlation function(CCF thereafter). Each row of this table displays the CCF between a given detrended fiscal vari-able at time t + k, and detrended GDP at time t. For the sake of robustness, we show results for aset of standard filters19 as applied to seasonally adjusted time series in real terms.

Each row of this table displays the CCF between a measure of detrended real GDP at time t,and a detrended fiscal variable at time t + k. Following the standard discussion in the literature,it is said that two variables co-move in the same direction over the cycle if the maximum valuein absolute terms of the estimated correlation coefficient of the detrended series (call it dominantcorrelation) is positive, that they co-move in opposite directions if it is negative, and that they donot co-move if it is close to zero. A cut-off point of 0.20 roughly corresponds in our sample tothe value required to reject at the 5% level of significance the null hypothesis that the populationcorrelation coefficient is zero. Finally, the fiscal variable is said to be lagging (leading) the privatesector variable if the maximum correlation coefficient is reached for negative (positive) values ofk.

The results in the table show the strong and pro-cyclical behavior of total government revenuein the euro area, which follows the business cycle behavior in upturns and downturns, reflectingthe operation of automatic stabilizers. In addition, public revenues are much more volatile thanGDP, more than 1.5 times, on average. This reflects the fact that a number of taxes, most notablycorporate taxes, property taxes and other indirect taxes, tend to follow boom–bust dynamics anddo react to the cycle more than proportionally (Morris & Schuknecht, 2007). Finally, it is worthmentioning that the dominant correlation is the contemporaneous one (zero lag), reflecting that taxreceipts are particularly endogenous with respect to the business cycle. For the whole sample thecorrelation is 0.75, while if the most recent crisis years are removed from the sample (“pre-crisissample”) the dominant correlation is lower of 0.61, reflecting that the 2008–2012 has increasedthe synchronicity of public revenues and real GDP.

Given the not-quite-surprising feature that government revenues are strongly pro-cyclical, moststudies look at the cyclical properties of government spending (see Frankel, Vegh, and Vuletin

19 The selected filters are: (i) first difference filter; (ii) linear trend; (iii) Hodrick–Prescott filter for two alternative valuesof the band-pass parameter (the standard 1600, that is a fair approximation of the cycles of France and Italy, while a highervalue would be more appropriate for countries with more volatile cycles like Spain, as shown by Marcet and Ravn (2004));and (iv) Band-Pass filter (with two different band-pass parameters, capturing fluctuations between 1.5 and 8 years andbetween 1.5 and 12 years, an observation closer to average euro area business cycle duration).


Please cite

this article

in press

as: Paredes,

J., et

al. Fiscal

policy analysis

in the

euro area:

Expanding

thetoolkit.

Journal of

Policy M

odeling (2014),


AR

TIC

LE

IN P

RE

SS

+Model

JPO-6146;

N

o. of

Pages 24

16

J. Paredes

et al.

/ Journal

of Policy

Modeling

xxx (2014)

xxx–xxxTable 3Stylized facts on the cyclical properties of euro area aggregate total government revenues and total government expenditures: whole sample (1980Q1–2021Q4), pre-crisis sample(1980Q1–2007Q4) and euro area sample (1992Q1–2012Q4, including the Maastricht, run-up to EMU period).

Relative standarddeviation

CCF (real GDPt, Ft+k) Dominant correlation

k

−6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6

F = Total revenueWhole sample (1980Q1–2012Q4)First diff. filter 1.8 −0.05 0.11 0.13 0.14 0.24 0.30 0.47 0.40 0.23 0.23 0.09 0.01 0.10 → pro-cycl., contemp.Linear trend 1.1 0.26 0.39 0.51 0.61 0.71 0.78 0.82 0.81 0.76 0.69 0.59 0.48 0.38 → pro-cycl., contemp.HP 1600 1.7 −0.07 0.10 0.27 0.43 0.59 0.71 0.78 0.75 0.64 0.50 0.33 0.16 0.03 → pro-cycl., contemp.HP 3200 1.5 0.05 0.21 0.37 0.52 0.65 0.76 0.82 0.79 0.70 0.58 0.43 0.28 0.15 → pro-cycl., contemp.BP (1.5, 8 years) 1.9 −0.18 −0.01 0.19 0.40 0.59 0.72 0.78 0.75 0.63 0.46 0.27 0.07 −0.11 → pro-cycl., contemp.BP (1.5, 12 years) 1.6 0.09 0.24 0.41 0.57 0.71 0.80 0.84 0.81 0.72 0.59 0.43 0.27 0.11 → pro-cycl., contemp.

Average 1.6 0.02 0.17 0.31 0.45 0.58 0.68 0.75 0.72 0.61 0.51 0.36 0.21 0.11 → pro-cycl., contemp.

Pre-crisis sample (1980Q1–2007Q4)Average 1.9 0.15 0.30 0.36 0.42 0.49 0.53 0.61 0.60 0.54 0.53 0.45 0.38 0.34 → pro-cycl., contemp.

Euro area sample (1992Q1–2012Q4)Average 1.7 0.01 0.15 0.31 0.49 0.63 0.73 0.79 0.72 0.58 0.40 0.21 0.04 −0.10 → pro-cycl., contemp.

F = Total expenditureWhole sample (1980Q1–2012Q4)First diff. filter 0.6 0.01 0.05 −0.04 −0.05 0.04 −0.04 −0.01 0.02 −0.05 0.03 0.16 0.28 0.22 → pro-cycl., laggedLinear trend 1.0 −0.39 −0.32 −0.25 −0.18 −0.10 −0.02 0.06 0.12 0.18 0.25 0.32 0.38 0.43 → pro-cycl., laggedHP 1600 0.8 −0.25 −0.24 −0.24 −0.24 −0.22 −0.21 −0.17 −0.11 −0.03 0.10 0.25 0.39 0.50 → pro-cycl., laggedHP 3200 0.9 −0.34 −0.30 −0.26 −0.21 −0.16 −0.10 −0.02 0.06 0.16 0.28 0.41 0.53 0.63 → pro-cycl., laggedBP (1.5, 8 years) 0.6 −0.10 −0.12 −0.17 −0.22 −0.27 −0.30 −0.31 −0.28 −0.20 −0.09 0.07 0.24 0.39 → pro-cycl., laggedBP (1.5, 12 years) 0.9 −0.45 −0.41 −0.36 −0.31 −0.25 −0.18 −0.09 0.01 0.12 0.26 0.39 0.53 0.65 → pro-cycl., lagged

Average 0.8 −0.25 −0.22 −0.22 −0.20 −0.16 −0.14 −0.09 −0.03 0.03 0.14 0.27 0.39 0.47 → pro-cycl., lagged

Pre-crisis sample (1980Q1–2007Q4)Average 1.1 −0.34 −0.25 −0.21 −0.14 −0.02 0.08 0.23 0.36 0.44 0.55 0.62 0.67 0.64 → pro-cycl., lagged

Euro area sample (1992Q1–2012Q4)Average 0.5 −0.15 −0.15 −0.18 −0.20 −0.24 −0.28 −0.29 −0.28 −0.21 −0.10 0.08 0.26 0.40 → pro-cycl., lagged

Note: Nominal fiscal variables are deflated using AWM’s GDP deflator. Quarterly real GDP is also taken from the latter database.





(2011), and the references quoted therein). Indeed, an important reason for the usual finding of pro-cyclical spending is precisely that government receipts get increased in booms, typically beyondexpectations (see discussion on revenue windfalls above), and thus governments use the surplus toincrease spending proportionately as a consequence of political pressure or just following certainsocial-welfare-improving objectives.

As expected, in Table 3 total expenditure appears pro-cyclical as well, but lagged, inline withavailable evidence for the euro area obtained with annual data (see Lamo et al. (2013)); thisbehavior can be rationalized on the basis of the political economy arguments mentioned in theprevious paragraph. The lag detected with quarterly data implies that total expenditure followsGDP with a delay of 1 to 1½ years. Budgetary patterns on the spending side tend to be quitepersistent, in particular as regards sizeable items like public wages or public employment. Forexample, only in the period following an economic downturn are fiscal consolidation measuresimplemented, while in expansions, fresh government revenues tend to expand the public sectorwage bill with some delay. When comparing the three analyzed samples, it seems that the crisisand subsequent fiscal consolidation period (2008–2012) reduced the pro-cyclical correlation ofpublic spending. While for the whole sample the dominant correlation stands at about 0.5, forthe pre-crisis period it was significantly higher, of the order of 0.7. This may reflect the factthat, firstly, the 2008–2009 crisis did not affect a number of current spending items as much asin other similar periods in the past, while, secondly, and on different grounds, the most recent(and ongoing) fiscal consolidation period may have somewhat weakened the traditional politicaleconomy channels outlined above, given sizeable consolidation needs.

The three main components of public spending (Table 4), namely government consumption,social payments and government investment, reflect the same pro-cyclical pattern than totalspending, overall. Interestingly, the estimated cyclical pattern and correlation of consumptionexpenditure was not significantly affected by the past five years (dominant correlation of 0.58 forthe whole sample, and 0.64 when 2008–2012 is excluded). Thus, the weakening of the correlationis a reflection of the impact of the most recent data on social payments and government investment.

As regards social payments other than unemployment benefits, the weak pro-cyclical, laggedpattern of 0.34 estimated with the pre-crisis sample, gets reduced when the whole sample is used;what is more, when the 1980s and the first part of the 1990s are excluded, the overall patternis a weak, counter-cyclical one. Again, in this respect, the containment of social transfers likepensions over the crisis and the subsequent period of fiscal prudence/consolidation may explainthe change in the (in any case weak) cyclical pattern. Regarding the most volatile componentof social payments, namely unemployment expenditures, they became more responsive to thecycle, as the negative correlation (counter-cyclical pattern) was of −0.49 for the whole samplebut turned out to increase (in absolute value) to −0.79 when the 1992Q1–2012Q4 period isconsidered. Unemployment-related benefits increase, as expected, in downturns and decrease inupturns; at the same time, unemployment spending seems to lead real GDP by 1–2 quarters. Thelatter evidence is consistent with an interpretation whereby employment losses at the beginning ofa cyclical downturn tend to be associated with new unemployed receiving full-entitlement benefits(given that downturns do occur after a good times period), coupled with the fact that the averageduration of the entitlement tends to be lower than the number of quarters the economy is belowtrend.

Finally, government investment presents a pro-cyclical and lagged behavior, with a dominantcorrelation of 0.48 for the whole sample, a number lower than the correlation estimated for thepre-crisis sample of 0.59. The lessening of the pro-cyclical bias in investment expenditure reflectsthe fact that the fiscal consolidation process that started by 2010 hinged heavily, and in a quite


Please cite

this article

in press

as: Paredes,

J., et

al. Fiscal

policy analysis

in the

euro area:

Expanding

thetoolkit.

Journal of

Policy M

odeling (2014),


AR

TIC

LE

IN P

RE

SS

+Model

JPO-6146;

N

o. of

Pages 24

18

J. Paredes

et al.

/ Journal

of Policy

Modeling

xxx (2014)

xxx–xxxTable 4Stylized facts on the cyclical properties of the main components of euro area fiscal expenditures: whole sample (1980Q1–2021Q4), pre-crisis sample (1980Q1–2007Q4) andeuro area sample (1992Q1–2012Q4, including the Maastricht, run-up to EMU period).

Relative standarddeviation

CCF (real GDPt, Ft+k) Dominant correlation

k

−6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6

F = Government consumptionWhole sample (1980Q1–2012Q4)Average 0.9 −0.13 −0.14 −0.15 −0.13 −0.08 −0.04 0.03 0.08 0.17 0.29 0.40 0.49 0.58 → pro-cycl., laggedPre-crisis sample (1980Q1–2007Q4)Average 1.3 −0.21 −0.15 −0.11 −0.02 0.08 0.17 0.29 0.35 0.44 0.54 0.60 0.63 0.64 → pro-cycl., laggedEuro area sample (1992Q1–2012Q4)Average 0.8 −0.09 −0.12 −0.15 −0.18 −0.19 −0.18 −0.12 −0.10 0.01 0.14 0.27 0.40 0.52 → pro-cycl., lagged

F = Social payments other than unemployment benefitsWhole sample (1980Q1-2012Q4)Average 0.7 −0.09 −0.10 −0.13 −0.14 −0.13 −0.12 −0.08 −0.03 0.02 0.09 0.15 0.23 0.27 → pro-cycl., laggedPre-crisis sample (1980Q1–2007Q4)Average 1.2 −0.07 −0.04 −0.03 −0.01 0.05 0.10 0.16 0.23 0.28 0.32 0.32 0.34 0.32 → pro-cycl., laggedEuro area sample (1992Q1–2012Q4)Average 0.4 −0.25 −0.29 −0.32 −0.34 −0.35 −0.35 −0.31 −0.27 −0.21 −0.12 −0.03 0.11 0.21 → counter-cycl., lead.

F = Unemployment benefitsWhole sample (1980Q1–2012Q4)Average 37.0 −0.03 −0.10 −0.17 −0.27 −0.37 −0.44 −0.49 −0.48 −0.40 −0.26 −0.10 0.05 0.18 → counter-cycl.Pre-crisis sample (1980Q1–2007Q4)Average 60.1 −0.31 −0.33 −0.37 −0.40 −0.42 −0.44 −0.39 −0.32 −0.22 −0.08 0.05 0.17 0.28 → counter-cycl., lead.Euro area sample (1992Q1–2012Q4)Average 23.2 0.38 0.25 0.10 −0.12 −0.36 −0.54 −0.72 −0.79 −0.71 −0.54 −0.30 −0.09 0.08 → counter-cycl.

F = Government investmentWhole sample (1980Q1–2012Q4)Average 12.7 −0.04 −0.01 0.07 0.18 0.25 0.26 0.33 0.37 0.38 0.41 0.48 0.47 0.45 → pro-cycl., laggedPre-crisis sample (1980Q1–2007Q4)Average 18.7 0.08 0.14 0.22 0.38 0.48 0.50 0.58 0.59 0.53 0.51 0.52 0.45 0.38 → pro-cycl., laggedEuro area sample (1992Q1–2012Q4Average 13.3 −0.10 −0.05 0.05 0.11 0.15 0.17 0.22 0.22 0.26 0.34 0.40 0.41 0.42 → pro-cycl., lagged

Note: Nominal fiscal variables are deflated using AWM’s GDP deflator. Quarterly real GDP is also taken from the latter database.





persistent fashion, on this budgetary item, as discussed above, and as a consequence traditionalboom–bust dynamics were broken over the past few years.

5. Conclusions

The fiscal database developed in this paper presents the potential of constituting (and hasalready constituted) a useful input for broader macroeconomic and policy analyses using euro areadata and involving fiscal variables. Before the production of the current dataset those exerciseswere mainly conducted either with annual data or with limited availability of quarterly fiscalinformation. This type of studies have recently received renewed attention and include simulationexercises to assess the impact fiscal stimulus and fiscal adjustment packages, analyses of theinteraction between monetary and fiscal policies, or the estimation of fiscal policy rules.

As an illustration along these lines, we provide two types of empirical evidence. First, wecharacterize the evolution of the main euro area fiscal aggregates over the past four decades. Inthe case of the fiscal consolidation episode of 2010–2012, we show how it was basically hinging ontax increases and sizeable government investment cuts, a composition that cannot be consideredto be inline with the best policy strategies, as advocated by the literature. Secondly, we providestylized facts on the cyclical behavior of fiscal policies in the euro area. Here the main highlight isthat the headline public spending euro area aggregates behave in a pro-cyclical fashion, and followGDP with a delay of 1 to 1½ years. At the same time we provide evidence that this pro-cyclicalbias has weakened in recent years.

Appendix A. Input database

The bulk of annual euro area data in ESA95 terms for the period 1995–2012 is taken fromAMECO, the database of the Directorate-General for Economic and Financial Affairs of theEuropean Commission. There are two exceptions to this source: the series for annual euro areadirect taxes on corporations for the period 1980–2008 was obtained from the OECD EconomicOutlook database, while the series for employers’ social contributions (for the period 1991–2012)was taken from Eurostat’s ESA95 database. For the prior period 1980–1994, we had to deal with thepresence of a break in accounting standards (ESA79–ESA95) and the German unification. In orderto obtain homogeneous levels for the whole period 1980–2012, we removed level discontinuitiesby applying backwards the growth rates of the series in ESA79 terms (that exclude East Germany)to the levels of the ESA95 series.

Quarterly figures for the euro area aggregate for the period 1999Q1–2012Q4 are taken fromEurostat. The impact of one-off proceeds from the allocation of mobile licenses (UMTS), thatsizeably distort some years, was removed from the relevant series. Quarterly and monthly fiscalvariables (indicators) for the biggest euro area economies, namely Germany, France, Italy, Spainand the Netherlands, are taken from Eurostat (available ESA95 series), several national sources,the Bank of International Settlements (BIS), and other sources, as described in Table 1. Whennecessary, country variables are set into euros using the official fixed euro conversion rates. Also,when necessary, German series were corrected for the impact of the Unification. For additionaldetails on some data sources of monthly/quarterly “indicator” series, the interested reader can alsoconsult Onorante et al. (2010) and ECB (2004). Finally, annual information in ESA79/ESA95definitions for the individual countries is taken from the AMECO database when needed, and





quarterly information following ESA95 standards from Eurostat, as mentioned above for the euroarea aggregate.20

Appendix B. Econometric approach

The basic model is of the Unobserved Components class known as the Basic Structural Model(Harvey, 1989), that decomposes a set of time series in unobserved though meaningful componentsfrom an economic point of view (mainly trend, seasonal and irregular). The exposition in thissubsection follows closely Harvey (1989), Pedregal and Young (2002) and Young and Pedregal(1999). The model is multivariate, and may be written as Eq. (1), where t is a time sub-indexmeasured in months (for models set up at the monthly frequency),[

zt

ut

]= Tt + St + et (1)

where [zt, ut]T, Tt, St and et denote the m-dimensional output time series (broken down into ascalar output, zt, and indicators, ut), trend, seasonal and irregular components, respectively. Eq.(1) is in fact a set of observation equations in a State Space system, which has to be completed bythe standard transition or state equations. The state equations qualify the dynamic behavior of thecomponents. In this particular case, the transition equations for models of the trend and seasonalcomponents are a Local Linear Trend and the Trigonometric Seasonal (see either Harvey (1989),or Pedregal and Young (2002), for details).

The mixture of frequencies, and the estimation of models at the quarterly frequency, impliescombining variables that at the quarterly frequency can be considered as stocks with those beingpure flows. An annual ESA95 series cast into the quarterly frequency is a set of missing obser-vations for the first three-quarters of the year and the observed value assigned to the last monthof each year. Theoretically the annual ESA95 series would be obtained from a quarterly ESA95series by summation of the 4 quarters of a year (Q1–Q4) had them been available. Model (1) then,has to be adapted to the fact that in the same model one variable is on an annual sampling interval,while others are sampled at a quarterly rate. This is the so-called temporal aggregation problem,which is relatively easy to handle in the State Space framework. The way the models are definedspecifically may be seen in Pedregal and Pérez (2010) and Leal, Pedregal, and Pérez (2011). Foreach specific variable considered in this study, models of type (1) are estimated. In each model,the variable {zt} corresponds to the target time series to be interpolated, composed of annualobservations for the period 1980–1998, and quarterly observations for the period 1999–2012. Thevector of indicator variables {ut}, in turn, comprises a set of variables with quarterly (for quarterlymodels) observations, typically (but not always) available for the full period 1980–2012.

Estimation of model (1) provides estimates for the missing values in {zt} (missing quarterlydata points) and estimates of the vector comprising the unobserved components that includethe estimated seasonal components. Thus, it is possible to compute model-consistent seasonallyadjusted interpolated series for the target variables {zt} just by subtraction of the correspondinglyestimated seasonal components from {zt}. For all euro area models the vector {zt} encompassesannual ESA95 euro area data for the period 1980–1998, and quarterly, non-seasonally adjusted,ESA95 data for the period 1999Q1–2012Q4. On the other hand, as it is clear from the description

20 The vintages of data used are the ones that were available in April 2013.





of data sources in Table 1, in some instances it was necessary to use more than one source ofintra-annual information in order to compute the indicator variable finally included in the euroarea model within the vector {ut}.21

A final remark on the dimensionality of the models is worth mentioning. In order to reducethe dimensionality of our models and somewhat avoid the “curse of dimensionality” we opted forvariable-by-variable models. By this we mean that, in all cases, {zt} encompasses just one timeseries (annual/quarterly), and {ut} the set of indicators corresponding to the latter variable, with amaximum of five indicators (one per country for each variable). The alternative would have beento run models in which {zt} would have included several variables, and thus {ut} would havebeen a matrix with indicators by blocks for each component of {zt}. Examples of other suitablemodels include a joint model for TOR and TOE, as in Pedregal and Pérez (2010), i.e. {zt} = {TOR,TOE}, a joint model for the revenue side of the governments accounts, i.e. {zt} = {TOR, DTX,SCT, TIN, OTOR}, or a joint model for the expenditure side, i.e. {zt} = {TOE, THN, GCN,GIN, INP, SIN, OTOE}. We preferred to use for interpolation purposes more parsimoniousmodels, and thus disregarded the alternative approach, quite valid in different frameworks (likeforecasting).

References

Aarle, B. van., Garretsen, H., & Gobbin, N. (2003). Monetary and fiscal transmission in the euro-area: Evidence from astructural VAR analysis. Journal of Economics & Business, 55, 609–638.

Afonso, A., & Sousa, R. M. (2009a). The macroeconomic effects of fiscal policy. ECB Working Paper Series No. 991Afonso, A., & Sousa, R. M. (2009b). The macroeconomic effects of fiscal policy in Portugal: A Bayesian SVAR analysis.

School of Economics and Management Working Papers No. 09/2009/DE/UECE.Alesina, A., & Ardagna, S. (2010). Large changes in fiscal policy: Taxes versus spending. In J. R. Brown (Ed.), Tax policy

and the economy (Vol. 24). University of Chicago Press.Anderson, H., Dungey, M., Osborn, D. R., & Vahid, F. (2007). Constructing historical euro area data. Australian National

University, Centre for Applied Macroeconomic Analysis. CAMA Working Papers No. 2007-18.Asimakopoulos, S., Paredes, J., & Warmedinger, T. (2013). Forecasting fiscal time series using mixed frequency data.

ECB Working Paper Series No. 1550.Auerbach, A. J., & Gorodnichenko, Y. (2012). Output spillovers from fiscal policy. Vox.EU, 10 December.Barrios, S., & Rizza, P. (2010). Unexpected changes in tax revenues and the stabilisation function of fiscal policy: Evidence

for the European Union, 1999–2008. European Commission Economic Papers No. 404.Batini, N., Callegari, G., & Melina, G. (2012). Successful Austerity in the United States, Europe and Japan. IMF Working

Papers WP/12/190.Baum, A., & Koester, G. B. (2011). The impact of fiscal policy on economic activity over the business cycle – evidence

from a threshold VAR analysis. In Discussion Paper Series 1: Economic Studies 2011,03. Deutsche Bundesbank,Research Centre.

Beetsma, R., & Giuliodori, M. (2011). The effects of government purchases shocks: Review and estimates for the EU.Economic Journal, 121, F4–F32.

Beetsma, R., Giuliodori, M., & Klaasen, F. (2006). Trade spill-overs of fiscal policy in the European Union: A panelanalysis. Economic Policy, 21, 640–687.

Bénassy-Quéré, A., & Cimadomo, J. (2012). Changing patterns of domestic and cross-border fiscal policy multipliers inEurope and the US. Journal of Macroeconomics, 34, 845–873.

Beyer, A., Doornok, J. A., & Hendry, D. F. (2001). Constructing historical euro-zone data. Economic Journal, 111,102–121.

Bi, H., Leeper, E. M., & Leith, C. (2013). Uncertain fiscal consolidations. Economic Journal, 123, F31–F63.

21 In a document available upon request we provide a quite detailed description of the implementation of the generalmethodology and the data inputs described in the case of each one of the variables included in our study, and also thedescription of the components of {ut} in each case.


http://refhub.elsevier.com/S0161-8938(14)00070-2/sbref0005












































































































































































































































































































































































Bosker, E. M. (2006). On the aggregation of eurozone data. Economics Letters, 90, 260–265.Brüggemann, R., & Lütkepohl, H. (2006). A small monetary system for the euro area based on German data. Journal of

Applied Econometrics, 21, 683–702.Bruneau, C., & de Bandt, O. (2003). Monetary and fiscal policy in the transition to EMU: What do SVAR models tell us?

Economic Modelling, 20, 959–985.Burriel, P., de Castro, F., Garrote, D., Gordo, E., Paredes, J., & Pérez, J. J. (2010). Fiscal policy shocks in the euro area

and the US: An empirical assessment. Fiscal Studies, 31, 251–285.Canova, F., & Pappa, E. (2011). Fiscal policy, pricing frictions and monetary accommodation. Economic Policy, 26,

555–598.Canzoneri, M. B., Cumby, R. E., & Diba, B. (2003). New views on the transatlantic transmission of fiscal policy and

macroeconomic policy coordination. In M. Buti (Ed.), Monetary and fiscal policies in EMU. Cambridge UniversityPress.

Cimadomo, J. (2011a). The fiscal stimulus and challenges ahead: Views on the euro area. Economic Papers, 30, 23–28.Cimadomo, J. (2011b). Fiscal policy in real-time. The Scandinavian Journal of Economics, 114, 440–465.Cimadomo, J., Kirchner, M., & Hauptmeier, S. (2010). Transmission of government spending shocks in the euro area:

Time variation and driving forces. ECB Working Paper Series No. 1219.Coenen, G., Christoffel, K., & Warne, A. (2008). The new area-wide model of the euro area: A micro-founded open-

economy model for forecasting and policy analysis. ECB Working Paper Series No. 944.Coenen, G., Straub, R., & Trabandt, M. (2012). Fiscal policy and the great recession in the euro area. American Economic

Review, Papers and Proceedings, 102, 71–76.Coenen, G., Straub, R., & Trabandt, M. (2013). Gauging the effects of fiscal stimulus packages in the euro area. Journal

of Economic Dynamics and Control, 37, 367–386.Cogan, J. F., Cwik, T., Taylor, J. B., & Wieland, V. (2009). New Keynesian versus old Keynesian government spending

multipliers. Journal of Economic Dynamics and Control, 34, 281–295.Corsetti, G., Meier, A., & Müller, G. J. (2010). Cross-border spillovers from fiscal stimulus. International Journal of

Central Banking, 6, 5–37.Creedy, J., & Gemmell, N. (2002). The built-in flexibility of income and consumption taxes. Journal of Economic Surveys,

14, 509–532.Creedy, J., & Gemmell, N. (2007). Corporation tax buoyancy and revenue elasticity in the UK. Economic Modelling, 25,

24–37.Cwik, T., & Wieland, V. (2010). Keynesian government spending multipliers and spillovers in the euro area. ECB Working

Paper No. 1267.Davig, T., & Leeper, E. (2011). Monetary–fiscal policy interactions and fiscal stimulus. European Economic Review, 55,

211–227.de Castro, F. (2006). The macroeconomic effects of fiscal policy in Spain. Applied Economics, 38, 913–924.de Castro, F., & Garrote, D. (2012). The effects of fiscal shocks on the exchange rate in the EMU and differences with the

US. Banco de Espana Working Papers 1224, Banco de Espana.de Castro, F., & Hernández de Cos, P. (2008). The economic effects of fiscal policy: The case of Spain. Journal of

Macroeconomics, 30, 1005–1028.Dilnot, A. (2012). Numbers and public policy: The power of official statistics and statistical communication in public

policymaking. Fiscal Studies, 33, 429–448.Draghi, M. (2013). Introductory statement to the press conference (with Q&A), Frankfurt am Main, 5 September 2013.Dreger, C., & Marcellino, M. (2007). A macroeconometric model for the euro economy. Journal of Policy Modeling, 29,

1–13.Duisenberg, W. F. (2003). Monetary and fiscal policy in the euro area. In Introduction at the international monetary

conference, Berlin, 3 June 2003. Available from. http://www.ecb.int/key/03/sp030603.htmlEuropean Central Bank. (2004, August). Properties and use of general government quarterly accounts. ECB Monthly

Bulletin.European Central Bank. (2008, July). One monetary and many fiscal policies: Ensuring a smooth functioning of EMU.

ECB Monthly Bulletin.European Central Bank. (2009, September). ECB Monthly Bulletin, 81–89. Section 5 (Boxes 9 and 10).European Commission (2002). Reg. (EC) No 1221/2002 of the European Parliament and of the Council of 10 June 2002

on quarterly non-financial accounts for general government.European Commission. (2002). ESA95 manual on government deficit and debt, methods and nomenclatures series.

Luxembourg: European Commission.
















































































































































































































































































































































































































































































































































































































http://www.ecb.int/key/03/sp030603.html







































































European Commission. (2006). Manual on quarterly non-financial accounts for general government. Luxembourg:European Commission. http://epp.eurostat.ec.europa.eu/cache/ITY SDDS/Annexes/gov q ggnfa esms an7.pdf

European Commission. (2007). Quarterly government accounts and fiscal surveillance in public finances in EMU.Brussels: European Commission.

European Commission. (2012). Report on Public finances in EMU. European economy 4/2012. Brussels: EuropeanCommission.

Fagan, G., Henry, J., & Mestre, R. (2001). An area-wide model (AWM) for the euro area. European Central Bank WorkingPaper Series No. 42.

Fagan, G., Henry, J., & Mestre, R. (2005). An area-wide model (AWM) for the euro area. Economic Modelling, 22, 39–59.Fatas, A., & Mihov, I. (2009). The euro and fiscal policy. NBER Working Paper No. 14722.Favero, C., & Marcellino, M. (2005). Modelling and forecasting fiscal variables for the euro area. Oxford Bulletin of

Economics and Statistics, 67, 755–783.Forni, L., Monteforte, L., & Sessa, L. (2009). The general equilibrium effects of fiscal policy: Estimates for the euro area.

Journal of Public Economics, 93, 559–585.Frankel, J., Vegh, C. A., & Vuletin, G. (2011). Fiscal policy in developing countries: Escape from procyclicality. VoxEU,

23 June 2011.Giordano, R., Momigliano, S., Neri, S., & Perotti, R. (2007). The effects of fiscal policy in Italy: Evidence from a VAR

model. European Journal of Political Economy, 23, 707–733.Harvey, A. (1989). Forecasting structural time series models and the Kalman filter. Cambridge University Press.Heppke-Falk, K. H., Tenhofen, J., & Wolff, G. B. (2006). The macroeconomic effects of exogenous fiscal policy shocks in

Germany: A disaggregated SVAR analysis. Deutsche Bundesbank Discussion Paper Series 1: Economic Studies No41/2006.

Jacobs, J. P. A. M., Kuper, G. H., & Sterken, E. (2003, November). A structural VAR model of the euro area. In CPBworkshop on economic policy modelling The Hague.

Jacobs, J. P. A. M., Kuper, G. H., & Verlinden, J. (2007). Monetary policy in the euro area: The impact of fiscal closurerules. Institute of Economics and Econometrics, University of Groningen. Working Paper 2007/01.

Kollmann, R., Ratto, M., Roeger, W., & in’t Veld, J. (2013). Fiscal policy, banks and the financial crisis. Journal ofEconomic Dynamics and Control, 37, 387–403.

Lamo, A., Pérez, J. J., & Schuknecht, L. (2013). The cyclicality of consumption, wages and employment of the publicsector in the euro area. Applied Economics, 45, 1551–1569.

Lane, P. (2003). The cyclical behaviour of fiscal policy: Evidence from the OECD. Journal of Public Economics, 87,2661–2675.

Leal, T., Pérez, J. J., Tujula, M., & Vidal, J. P. (2008). Fiscal forecasting: Lessons from the literature and challenges.Fiscal Studies, 29, 347–386.

Leal, T., Pedregal, D. J., & Pérez, J. J. (2011). Short-term monitoring of the Spanish government balance. SERIES-Journalof the Spanish Economic Association, 2, 97–119.

Marcellino, M. (2006). Some stylized facts on fiscal policy in the euro area. Journal of Macroeconomics, 28, 461–479.Marcet, A., & Ravn, M. (2004). The HP-filter in cross-country comparisons. CEPR DP 4244.Morris, R., & Schuknecht, L. (2007). Structural balances and revenue windfalls – The role of asset prices revisited. ECB

Working Paper No. 737.Morris, R., Bras, C. R., Castro, F. de, Jonk, S., Kremer, J., Linehan, S., et al. (2009). Explaining government revenue

windfalls and shortfalls: An analysis for selected EU countries. ECB Working Paper Series No. 1114.Onorante, L., Pedregal, D. J., Pérez, J. J., & Signorini, S. (2010). The usefulness of infra-annual government cash budgetary

data for fiscal forecasting in the euro area. Journal of Policy Modeling, 32, 98–119.Pappa, E. (2009). The effects of fiscal expansions: An international comparison. Barcelona GSE Working Paper No. 409.Pedregal, D. J., & Pérez, J. J. (2010). Should quarterly government finance statistics be used for fiscal surveillance in

Europe? International Journal of Forecasting, 26, 794–804.Pedregal, D. J., & Young, P. C. (2002). Statistical approaches to modelling and forecasting time series. In M. Clements,

& D. Hendry (Eds.), Companion to economic forecasting (pp. 69–104). Oxford: Blackwell Publishers.Penner, R. G. (2008). Federal revenue forecasting. In J. Sun, & T. D. Lynch (Eds.), Government budget forecasting:

Theory and practice. Auerbach Publications (Chapter 2).Pérez, J. J. (2007). Leading indicators for euro area government deficits. International Journal of Forecasting, 23, 259–275.Pérez, J. J., & Sánchez, A. J. (2011). Is there a signalling role for public wages? Evidence for the euro area based on

macro data. Empirical Economics, 41, 421–445.Perotti, R. (1996). Fiscal consolidation in Europe: Composition matters. American Economic Review, 86, 105–110.


http://epp.eurostat.ec.europa.eu/cache/ITY_SDDS/Annexes/gov_q_ggnfa_esms_an7.pdf













































































































































































































































































































































































































































































































































































































































































































Ratto, M., Roeger, W., & in’t Veld, J. (2009). Quest III, an estimated open-economy DSGE model of the Eurozone withfiscal and monetary policy. Economic Modelling, 26, 222–233.

Silvestrini, A., Salto, M., Moulin, L., & Veredas, D. (2008). Monitoring and forecasting annual public finance deficitevery month: The case of France. Empirical Economics, 34, 493–524.

Smets, F., & Wouters, R. (2003). An estimated dynamic stochastic general equilibrium model of the euro area. Journalof the European Economic Association, 1, 1123–1175.

Young, P. C., & Pedregal, D. J. (1999). Recursive and en-block approaches to signal extraction. Journal of AppliedStatistics, 26, 103–128.
































































































fiscal policy analysis in the euro area: expanding the toolkit

Documents