assessing the impact of fiscal measures on the czech economy

Economic Modelling xxx (2014) xxx–xxx

ECMODE-03406; No of Pages 8

Contents lists available at ScienceDirect

Economic Modelling

j ourna l homepage: www.e lsev ie r .com/ locate /ecmod

Assessing the impact of fiscal measures on the Czech economy

Róbert Ambriško a,c,⁎, Jan Babecký a, Jakub Ryšánek a,d, Vilém Valenta b

a Czech National Bank, Prague, Czech Republicb European Central Bank, Frankfurt, Germanyc CERGE-EI, a joint workplace of Charles University and the Economics Institute of the Academy of Sciences of the Czech Republic, Prague, Czech Republicd University of Economics, Prague, Czech Republic

⁎ Corresponding author.E-mail addresses: [email protected] (R. Ambrišk

(J. Babecký), [email protected] (J. Ryšánek), vilem.val

http://dx.doi.org/10.1016/j.econmod.2014.07.0210264-9993/© 2014 Elsevier B.V. All rights reserved.

Please cite this article as: Ambriško, R., et adx.doi.org/10.1016/j.econmod.2014.07.021

a b s t r a c t
a r t i c l e i n f o
Article history:Accepted 14 July 2014Available online xxxx

Keywords:Bayesian estimationDSGEFiscal multipliersFiscal policyFiscal ruleOpen economyUnemployment

Webuild a DSGEmodel to investigate the transmission of fiscal policy to the real economy in the Czech Republic.Departing from the elements of the CzechNational Bank's current g3 forecastingmodel (Andrle,Hlédik, Kameník,and Vlček, 2009), we introduce a comprehensive fiscal sector incorporating a number of modeling features thatare often neglected in the mainstream DSGE literature, e.g. allowing government consumption and governmentcapital to be productive. Furthermore, we extend our fiscal model to include unemployment in a way proposedby Galí, López-Salido, and Vallés (2007). Crucial fiscal parameters, related mainly to the specified fiscal rule, areestimated using Bayesian techniques. The model is then used to calculate a set of fiscal multipliers for individualrevenue and expenditure items of the government budget. We find that the largest real GDP fiscal multipliers inthe first year are associatedwith government consumption (0.6), social security contributions paid by employers(0.6), and government investment (0.5).

o), [email protected]@ecb.int (V. Valenta).

l., Assessing the impact of fiscal measures on

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Fiscal policy has received considerable attention since the globaleconomic and financial crisis began in 2008. This attention has also fall-en on central banks, since fiscal measures often significantly affect eco-nomic activity (real GDP, inflation) and consequently monetary policyinterest rates need to be set appropriately. Governments frequently in-troduce several fiscal measures at once, in so-called fiscal packages,which consist of variousmeasures on both the expenditure and revenuesides of the government budget. Depending on the model used, fiscalmeasures might not be easily implemented into the model to producemacroeconomic forecasts fully consistent with the fiscal policy settings.This applies also to the practice in the Czech National Bank (CNB).

The CNB's core g3 model, developed and described in the study ofAndrle et al. (2009), currently lacks sufficient detail with respect tothe fiscal sector. Hence, the objective of this paper is to address this de-ficiency by building a satellite DSGE model with an extended fiscal sec-tor. To concentrate mainly on fiscal policy variables, we proceed withsome simplification of the g3 model; to be more specific, stochastictrends are omitted from the model, but on the other hand we enrichthe model with several important fiscal channels. First, we introduceinto our model so-called “rule-of-thumb” households, in the mannerof Galí et al. (2007). These households do not accumulate any savingsand consume all their disposable income. Second, we allowgovernment

consumption and government capital to be productive; in other words,government consumption brings some utility to households and gov-ernment capital contributes to firms' production. These two productivefeatures are often neglected in the mainstream DSGE literature,although the possibility of productive government consumption andgovernment capital has been discussed in the past (Bailey, 1971;Barro, 1981; Baxter and King, 1993). Third, our model contains an ex-tensive set of fiscal instruments, namely, four instruments on the ex-penditure side (government consumption, government investment,unemployment benefits, and other social benefits) and five instrumentson the revenue side (a consumption tax, a wage tax, a capital tax, socialsecurity contributions paid by employers, and a lump-sum tax). Fourth,we extend our model to include unemployment, in a tractable way pro-posed by Galí (2011), which helps to partly endogenize unemploymentbenefits. Fifth, we specify the government's fiscal rule with feedback co-efficients for domestic output and debt, as found, for example, in Leeperet al. (2010), and estimate its coefficients for Czech data using Bayesiantechniques.

Given the absence of a consensus in the literature regarding the pre-cise value of the fiscal multipliers, one of our objectives is to provide theCNBwith the values of the fiscalmultipliers for the Czech economy. OurDSGE model with its extended fiscal sector allows us to produce a richset of multipliers, by several fiscal instrument categories, and also to as-sess the robustness of the multipliers to the underlying model assump-tions. Recent estimates offiscalmultipliers for the Czech economy basedon the structural VAR approach (Valenta, 2011) indicate an output fiscalmultiplier of between 0.3 and 0.6 in the first year after a shock togovernment spending. The real GDP fiscal multipliers implied by our

the Czech economy, Econ. Model. (2014), http://

http://dx.doi.org/10.1016/j.econmod.2014.07.021

mailto:[email protected]





http://www.sciencedirect.com/science/journal/02649993



1 This extension is needed to introduce unemployment into the model; for elaborationsee Galí (2011) or Ambriško et al. (2012).

2 R. Ambriško et al. / Economic Modelling xxx (2014) xxx–xxx

DSGE model attain their largest values in the first year in the case ofgovernment consumption (0.6), social security contributions paid byemployers (0.6), and government investment (0.5). These are followedby the fiscal multipliers for the consumption tax, the wage tax, and un-employment benefits (all roughly 0.2), then by other social benefits,lump-sum taxes, and the capital tax (0.1). Our results suggest that themost costly, in terms of the real GDP loss in the first year, are fiscalconsolidations based on cuts in government consumption and increasesin social security contributions paid by employers, followed by cuts ingovernment investment.

Comparing our estimates offiscalmultiplierswith the results report-ed in ameta analysis by Gechert andWill (2012) based on the examina-tion of 89 studies suggests that the rather low values of the fiscalmultipliers for the Czech economy could be attributed to its high importintensity of GDP. Furthermore, in what follows our DSGE-based fiscalmultipliers should be viewed as lower bound estimates compared tothose produced by macroeconometric models, single equation ap-proaches or VARs. Nevertheless, our sensitivity checks demonstratethat the higher the share of “rule-of-thumb” households, the higherthe values of fiscal multipliers, which corresponds to the evidencefrom the meta analysis.

For practical purposes, using our proposed DSGEmodel we evaluatethe partial impact of selected fiscal measures on the Czech economy(that is, not accounting for all supply-side effects and assuming that fis-cal shocks are of a temporary nature and the economy starts from itsequilibrium). We find that the selected fiscal consolidation measures,related to the ongoing process of fiscal consolidation in the CzechRepublic,might slow real GDPgrowth downby0.4, 0.8, and1.1 percent-age points in 2013, 2014, and 2015, respectively, as compared to thebaseline with unchanged fiscal policy.

The paper is structured as follows. Section 2 reviews relevant litera-ture, Section 3 outlines our satellite DSGE model with the extendedfiscal sector, and Section 4 provides estimates of fiscal multipliers andseveral robustness checks, and quantifies the impact of the selectedfiscal measures on the Czech economy. The last section summarizesour findings and outlines suggestions for future research.

2. Related literature review

Generally, the empirical literature provides a variety of estimates offiscalmultipliers. These are often based on (S)VAR techniques, but DSGEmodel estimates have gained in prominence recently. A number of com-prehensive models have been built, for instance the IMF's Global Inte-grated and Fiscal Model (see Kumhof et al., 2010), the EuropeanCommission's QUEST III model (Ratto et al., 2009), and the EuropeanCentral Bank's the New Area-Wide Model (Christoffel et al., 2008). Thelast-mentioned was recently extended by Coenen et al. (2012b) to in-clude a richer specification of the fiscal sector, and identified a signifi-cant role of discretionary fiscal policies for real GDP growth during theGreat Recession. An interesting comparison of structural models, interms of fiscal policy effectiveness, was performed by Coenen et al.(2012a). These authors found considerable agreement across modelson both the absolute and relative sizes of different types of fiscal multi-pliers. A recentmeta regression analysis of around90 studies byGechertandWill (2012) shows that the values of fiscalmultipliers are rather de-pendent on the chosen modeling approach and its settings; nonethe-less, the underlying studies suggest that the average fiscal multiplier isless than one.

Regarding the Czech Republic, there are several studies dealingwithfiscal policy. Barrell et al. (2004) examine the impacts of economicpolicies in several EU countries, including the Czech Republic, forwhich they estimate a fiscal multiplier of 0.4. A somewhat higher fiscalmultiplier of 0.6 is obtained by Hřebíček et al. (2005) using both regres-sion analysis and structural simulation. Recently, Prušvic (2010) deter-mines the government expenditure multiplier at 0.5, in line withprevious estimates. An extensive set of various fiscal multipliers for

Please cite this article as: Ambriško, R., et al., Assessing the impact of fidx.doi.org/10.1016/j.econmod.2014.07.021

the Czech Republic is provided by Klyuev and Snudden (2011), wherethe authors calibrate the IMF's GIMFmodel for the Czech data. The latestempirical evidence on the effects of fiscal policy on the Czech economyis provided by Valenta (2011) and Franta (2012). These studies exploreVAR-based identification approaches and apply classical (Valenta,2011) and Bayesian (Franta, 2012) estimation techniques. Accordingto Valenta (2011), the output fiscal multiplier is estimated to be in therange of 0.3–0.6 in the first year following a shock to governmentspending.

3. Structural DSGE model

Our structural model is built along the lines of themodels by Andrleet al. (2009), Coenen et al. (2012b), Galí (2011), and Galí et al. (2007).The small open economy is populated by two types of representativehouseholds, the first type called optimizers or Ricardian households,who can save, and the second type called “rule-of-thumb” consumersor non-Ricardian households, who cannot save and consume all theirdisposable income. The households consume a final consumptiongood, which is made from private consumption and government con-sumption goods. The members of households monopolistically supplya differentiated unit of labor to an employment agency, and wage set-ting follows Calvo contracts. Besides private capital, there is governmentcapital, which freely enters intermediate domestic goods production.Government expenditures are divided into government consumption,government investment, unemployment benefits, and other social ben-efits. Government revenues come from consumption, labor, capital, div-idend and lump-sum taxes, and social security contributions paid byemployers. The government balances its budget by issuing bonds orby adjusting taxes. In the fiscal rules, fiscal instruments (taxes or expen-ditures) react to the deviations of government debt and output fromtheir respective targets. The features of the model are shown in Fig. 1,where black parts overlap with the g3 model, red parts represent thefiscal sector, and green parts depict tax revenues. For a more detailedexposition of the model see Ambriško et al. (2012); in the followingtext we focus on the key equations.

The economy is populated by a continuumof households indexed byh ∈ [0, 1], of which rule-of-thumb households have a share γ andRicardian households 1 − γ. Each household has a continuum ofmembers1 indexed by a pair (i, j) ∈ [0, 1] × [0, 1], where index i standsfor the labor type and index j determines the disutility of work, specifiedas jϕn when themember is employed and zero otherwise, where ϕn ≥ 0is the elasticity of the marginal disutility of work. Both types ofhouseholds maximize their lifetime utility function given by:

EoX∞t¼o

βtUkh;t ¼

¼ EoX∞t¼o

βt�log Ck

h;t−χkt Ct−1

� �−θ

Z 1

0

Z Lkt ið Þ

0jϕn djdi

�¼

¼ EoX∞t¼o

βt log Ckh;t−χk

t Ct−1

� �− θ

1þ ϕn

Z 1

0Lkt ið Þ1þϕn di

� �ð1Þ

where β is the discount factor, superscript k ∈ {r, o} distinguishes rule-of-thumb and optimizer households, Ch,tk is the household-specificconsumption aggregate, Ct-1 is the lagged economy-wide level ofconsumption, Ltk(i) ∈ [0, 1] is the fraction of members of type i whoare employed in households of type k, θ is a parameter associated withthe disutility of labor supply, andχt

k is an exogenous processwhichdeter-mines the degree of internal habit formation. Households' consumption ismade up of private and government consumption goods as follows:

Ckt ¼ αCð Þ 1

vC Cpkt

� �vC−1vC þ 1−αCð Þ 1

vC Gkt

� �vC−1vC

� � vCvC−1

; ð2Þ

scal measures on the Czech economy, Econ. Model. (2014), http://



Fig. 1. The scheme of the model. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3R. Ambriško et al. / Economic Modelling xxx (2014) xxx–xxx

where αC is the share of the private good in the consumption aggregate,and vC N 0 is the elasticity of substitution between the private and govern-ment consumption good. The government good is equally available to allhouseholds and is provided free of charge. Total capital in the economy isthe aggregate of private and exogenously given government capital:

Kt ¼ αKð Þ 1vK Kp

t

� �vK−1vK þ 1−αKð Þ 1

vK Kgt

� �vK−1vK

h i vKvK−1

; ð3Þ

whereαK is the share of private capital in the capital aggregate, and vK N 0is the elasticity of substitution between private and government capital.

The households of optimizers respect the following budget constraint:

1þ τCt� �

PCt C

pot þ PI

tIpot þ Bo

t ≤ 1−τWt þ τUBt� �Z 1

0Wt ið ÞLot ið Þdiþ

þ 1−τKt� �

PKt þ τKt δ

pPIt

h iKpot−1 þ

þRt−1Bot−1 þ PC

t OBot−PC

t Tot þ 1−τDt

� �Dot ;

ð4Þ

where Cto is the optimizers' consumption; Itpo denotes optimizers' invest-

ment in private capital Ktpo; PtC, PtI are the unit prices of consumption

and investment goods; PtK is the rental rate of capital; Rt is the domesticnominal gross interest rate; Wt(i), Lt(i) are the nominal wage and opti-mizers' hours worked for labor of type i; τtC, τtW, τtK, τtD are consumption,wage, capital, and dividend taxes; τtUB is the unemployment benefitrate; OBto are the optimizers' other social benefits; δp is the depreciationrate of private capital; Bto are nominal domestic bonds issued by the gov-ernment andheld by optimizers; and Tt

o,Dto are the optimizers' lump-sum

taxes and dividends from monopolistic firms. The depreciation of capitalis exempted from capital tax.

Rule-of-thumb households spend their entire budget on consumption:

1þ τCt� �

PCt C

prt ≤ 1−τWt þ τUBt

� �Z 1

0Wt ið ÞLrt ið Þdiþ

þPCt OB

rt−PC

t Trt ;

ð5Þ

where Ctr, Ltr(i), OBtr, Ttr are the rule-of-thumbs' consumption, hours

worked for labor of type i, other social benefits, and lump-sum taxes.Government expenditures comprise government consumption, gov-

ernment investment, unemployment benefits and other social benefits


provided to households, and interest payments paid on issued debt. Thegovernment can issue bonds to finance its expenditures. Governmentrevenues are made up of consumption, labor, capital, dividend andlump-sum taxes, and social security contributions paid by employers.The total government budget balance can be computed by subtractinggovernment expenditures from government revenues:

BBt ¼ τCt PCt C

pt þ τWt þ τSt

� �WtLt þ τKt P K

t −δpP It

� �Kpt−1þ

þ τDt Dt þ PCt Tt−PG

t Gt−PItIgt−τUBt WtLtþ

−PCt OBt− Rt−1−1ð ÞBt−1;

ð6Þ

where Itg is government investment, τtS is the implicit tax rate for

social security contributions paid by employers, and Bt, OBt, Tt, Dt

are economy-wide aggregates of previously listed variables. Thegovernment's budget constraint follows: Bt-1 − BBt = Bt. Notethat in equilibrium the level of government debt is stable and thegovernment's budget is balanced.

Allfiscal instruments react to deviations of output and real debt fromtheir steady states. The only exceptions are the consumption tax, whichhas no feedback to output or debt, and the unemployment benefit rate,which responds only to deviations of the unemployment rate from itssteady state. Allowing for feedback effects, fiscal instruments can actprocyclically or countercyclically on the economy. Furthermore, on therevenue side shocks affecting one tax can affect other taxes as well.The motivation for such shocks is that the government often adjuststaxes jointly. For illustration purposeswe list two fiscal rules for govern-ment expenditures and one for tax revenues:

Gt

G¼ Yt

Y

� −ϕyg btb

� −ϕbg

ugt

τWtτW

¼ Yt

Y

� ϕytw btb

� ϕbtw

utct

� �ϕwc utkt

� �ϕkw utst

� �ϕsw utdt

� �ϕdwutwt

τUBtτUB

¼ ut

u

� �ϕuuubt

ð7Þ

where for x ∈ {g, tw}: ϕyx, ϕbx, ϕu are the feedback coefficients for out-put, debt, and the unemployment rate. If ϕyx is positive (negative),then a given fiscal instrument has a countercyclical (procyclical)





component. The shocks uxt ¼ ux

t−1� �ρx exp εxt

� �are serially correlated,

ρx ∈ [0, 1), and εtx are normal innovations. The remaining “cross-tax”co-movement coefficients ϕ[..] (powering the ut

x shocks) measure howmuch an unexpected movement in one tax affects another tax.

2 This MATLAB-based toolbox is a suitable framework for Bayesian estimations. For fur-ther information see www.dynare.org.

3 The selected mean of the distribution can be justified by a recent Gallup poll, where40% of approximately 1000 Czechs questioned said that they did not expect to make endsmeet (Ipsos Tambor, 2012). A roughly similar share of 37%was used by Štork and Závacká(2010) in the calibration of their model.

3.1. Calibration and estimation

As our model contains quite a large number of parameters, it wouldbe a tedious task to estimate all of them at once, so we intend to esti-mate only some of them for two reasons. First, some of ourmodel struc-tures overlap with the CNB's g3 model, which can serve as a referencepoint for calibration of selected parameters. Second, convergence ofthe estimation is of course a function of themagnitude of the parameterspace, and thus to avoid the curse of dimensionality we concentrateonly on a subset of parameters. Having a rich data set of various fiscalvariables, we estimate parameters in the fiscal feedback rules that aremost relevant to us in this study.

The share of the private good in the consumption good αC andthe share of private capital in the capital composite αK equal 0.8. Theparameters of the fiscal rules were estimated. However, some of thecoefficients — specifically the co-movement coefficient between con-sumption and capital taxes and the co-movement coefficients for thedividend tax — turned out to be insignificantly different from zero, sothese were set to zero. On the revenue side, the model works with im-plicit (or effective) tax rates, and their steady states were set as follows:the consumption tax at 20%, the wage tax at 10%, the capital tax at 19%,and social security contributions paid by employers at 24%. The divi-dend tax was turned off, because the impulse response functions forthe dividend shock lacked economic intuition; to be more specific, anincrease in the dividend tax led to an increase in real GDP. The steadystate value for the unemployment rate is set to 6.5%, which is thelong-run average for the Czech Republic. The steady state ratio of gov-ernment consumption to intermediate output was set to 25%, theshare of government investment in output equals 3%, unemploymentbenefits represents 0.3% of output, other social benefits make up 14%of output, and the debt (bonds) is calibrated to 60% of output. Further-more, we allow the steady state consumption of the two types of house-holds to differ, with the consumption of optimizers being higher thanthe consumption of rule-of-thumb households, reflecting the ideathat optimizers are wealthier than rule-of-thumb households. Specifi-cally, Co=Cr ¼ 1:25, the value also used by Coenen et al. (2011). Thedesired level of the steady state consumption ratio is delivered byadjusting lump-sum taxes for rule-of-thumb households in the steadystate. In our model, the actual steady state lump-sum taxes for rule-of-thumb households are negative, which means that rule-of-thumbhouseholds are subsidized by lump-sum transfers in the equilibrium.

Many of the modeled variables have reasonable counterparts in realdata, so it is possible to add a block of measurement equations and in-spect the solution of the model via Kalman filtering. This techniquealso provides the value of the logarithm of the likelihood function fora given parametrization of the model, which is useful when estimatingthe model parameters using Bayesian methods. The parameters inautoregressive processes measuring the persistence and volatility areestimated, where possible, equation by equation on real data.

The model is estimated on a total of 25 variables for the period from1996 to 2011 at quarterly frequency. Specifically, the data set covers theGDPexpenditure components (private consumption& investment, gov-ernment consumption & investment, exports, imports), including bothreal variables and their respective deflators, domestic financial variables(3-month PRIBOR rate, nominal CZK/EUR exchange rate), labor marketvariables (nominal wages, unemployment rate), a block of foreign vari-ables (3-month EURIBOR rate, GDP and PPI for EMU), and a rich set offiscal variables (consumption, capital and labor implicit tax rate, socialbenefits, unemployment benefits, primary budget balance, governmentdebt). Themajority of the data were collected from the Czech Statistical


Office, while domestic financial variables are from the CNB, implicit taxrates are OECD data, and foreign variables come from Eurostat.

Estimation of the model parameters is carried out in the DynareToolbox.2 The prior distributions of a subset of the model parametersget combined with the likelihood function based on the observed data.This results in the posterior distributions for particular parameters.First, we instruct Dynare to use numerical optimization techniques tosearch for the posterior modes. Next, we draw from the posterior distri-butions around thesemodes with help of the randomwalkMetropolis–Hastings (MH) algorithm. To make sure that convergence of theposterior simulations has been achieved, we run two parallel MH blocksof length 200,000 draws. Both simulations result in acceptance rates ofaround 38%. Half of the draws get thrown away as a burn-in.

The prior distributions for the estimated parameters were chosen asfollows. We assume the beta distribution for the share of rule-of-thumbhouseholds, centered on the value 0.4.3 Due to the zero lower bound,weimpose an inverse gammadistribution for the standard deviations of theshocks. For the remainder of the parameters the normal distribution isassumed. Along the lines of Galí (2011) we calibrate the mean valueof the elasticity of the marginal disutility of work ϕn at around 2. Theelasticities of substitution between the private and government inputsin the CES aggregates for consumption and capital, υC and nK, bothhave prior means calibrated to a value of 1. Themean value of the infla-tion feedback coefficient in the monetary policy rule is set at around 2.The prior means of the feedback coefficients in the fiscal rules are cali-brated to 0.25. The parameters thatmeasure the pair co-movements be-tween the fiscal rules should be positive by definition, as we think thatan increase in one tax is not systematically connected to a decrease ofanother tax. Therefore, the prior mean values of the co-movement pa-rameters are set to 0.25. We also have a strong belief that unemploy-ment benefits are positively correlated with the unemployment rate,thus the prior mean of ϕu is positive (0.25). The prior distributions ofthe standard deviations of the shock disturbances are all concentratedclose to the zero bound, as we keep their means at 0.01.

The estimation results are given in Table 1. As for the posterior out-comes from the Bayesian routine, the mean of the share of rule-of-thumb households is estimated at 16%. The inverse Frisch elasticity esti-mation exceeded our prior calibration, pointing to a lower elasticity ofhours worked to the wage rate. Compared to the prior setting, thelower posterior value of υC induces complementarity between theinput factors in the consumption CES aggregate. This is similar for υK, al-beit to a smaller extent. The posterior mean of ϕπ, the feedback coeffi-cient on inflation in the monetary policy rule, turned smaller than itsprior mean, reaching a value of 1.2. Not all the feedback coefficients inthe fiscal rules came out significantly different from zero in the estima-tion, so we keep some of them turned off. On the other hand, the dataprovide strong information on the coefficients capturing the reactionof fiscal measures to deviations of the gaps in the economy's intermedi-ate output and government indebtedness. The posterior means of theoutput feedback coefficients are positive, meaning that the respectivefiscal instruments act countercyclically (e.g. a positive output gap is con-nected with higher taxation or lower government expenditure). Theposterior means of the debt feedback coefficients are positive, helpingto stabilize government debt outside of equilibrium. The posteriormean of the unemployment feedback coefficient turned out to be slight-ly positive, and therefore unemployment benefits act countercyclicallyon the economy as well. (See Table 1.)

In order to assess whether the incorporation of a rich fiscal sectorinto the model structure improves the overall performance of our


http://www.dynare.org



Table 1Estimated parameters.

Parameter Prior distribution Posterior distribution

Mode Mean 10% 90%

Share of rule-of-thumb householdsγ B (0.4,0.1) 0.16 0.16 0.11 0.21

Inverse of Frisch elasticityϕn N (2.5,0.2) 2.92 2.95 2.74 3.17

Elasticities in CES aggregatesvC N (1,0.2) 0.75 0.64 0.23 1.03vK N (1,0.2) 0.95 0.97 0.70 1.24

Inflation feedback coefficientϕπ N (2,0.2) 1.20 1.19 0.92 1.46

Output feedback coefficientsϕyg N (0.25,0.1) 0.21 0.22 0.09 0.35ϕyig N (0.25,0.1) 0.22 0.22 0.09 0.35ϕyob N (0.25,0.1) 0.20 0.20 0.08 0.32ϕytk N (0.25,0.1) 0.17 0.17 0.04 0.30ϕytw N (0.25,0.1) 0.38 0.38 0.26 0.51ϕyts N (0.25,0.1) 0.28 0.27 0.14 0.40ϕytt N (0.25,0.1) 0.21 0.21 0.08 0.34

Debt feedback coefficientsϕbg N (0.25,0.1) 0.37 0.37 0.26 0.48ϕbig N (0.25,0.1) 0.32 0.32 0.19 0.44ϕbob N (0.25,0.1) 0.34 0.34 0.25 0.44ϕbtk N (0.25,0.1) 0.36 0.36 0.23 0.49ϕbtw N (0.25,0.1) 0.20 0.20 0.08 0.32ϕbts N (0.25,0.1) 0.20 0.22 0.09 0.35ϕbtt N (0.25,0.1) 0.36 0.38 0.26 0.50

Unemployment feedback coefficientϕu N (0.25,0.1) 0.06 0.06 0.02 0.10

Co-movement coefficientsϕwc N (0.25,0.1) 0.09 0.08 −0.04 0.22ϕsc N (0.25,0.1) −0.09 −0.07 −0.17 0.03ϕkw N (0.25,0.1) 0.26 0.25 0.15 0.36ϕsk N (0.25,0.1) −0.08 −0.07 −0.14 0.01ϕsw N (0.25,0.1) 0.26 0.25 0.12 0.38

Table 2Fiscal multipliers, baseline results.

Quarters Peak LR

1 4 8 16

One-year stimulusExpenditures (+)

Gov. consumption 0.66 0.63 0.58 0.46 0.66 0.52Gov. investment 0.45 0.45 0.47 0.45 0.48 0.34Unemployment benefits 0.14 0.19 0.29 0.30 0.30 0.28Other social benefits 0.12 0.12 0.14 0.11 0.22 0.22

Taxes (−)Consumption tax 0.18 0.24 0.28 0.16 0.28 0.02Wage tax 0.16 0.23 0.34 0.32 0.34 0.30Social contr. employers 0.21 0.59 0.75 0.42 0.76 0.42Capital tax 0.02 0.06 0.10 0.00 0.10 0.03Lump-sum tax 0.11 0.11 0.13 0.04 0.13 0.00

10-year stimulusExpenditures (+)

Gov. consumption 0.66 0.63 0.60 0.46 0.66 0.36Gov. investment 0.45 0.45 0.46 0.32 0.46 0.11Unemployment benefits 0.14 0.19 0.26 0.17 0.26 0.13Other social benefits 0.12 0.12 0.13 −0.07 0.13 −0.19

Taxes (−)Consumption tax 0.18 0.24 0.27 0.12 0.27 −0.07Wage tax 0.16 0.23 0.32 0.28 0.33 0.33Social contr. employers 0.21 0.59 0.72 0.43 0.72 0.40Capital tax 0.02 0.06 0.10 −0.07 0.10 −0.04Lump-sum tax 0.11 0.11 0.13 −0.06 0.13 −0.19

Note: LRmeans long-run. These are cumulative net-present-value fiscal multipliers calcu-lated as the discounted cumulative change in real GDP over the discounted cumulativechange in the corresponding fiscal instrument in nominal terms. The ex-ante fiscal stimu-lus lasts for one/ten year(s) and is calibrated so that the budget balance worsens by 1% ofnominal GDP in the first year.


model, we also estimated a version of the model which contains only alimited fiscal block. Turning off the fiscal sector entirely is not an option,because the model would become unstable. Instead, we minimize theeffects of the fiscal block as follows. We calibrate the shares αC, αK tounity, so government consumption and government capital are unpro-ductive. Further, the comprehensive fiscal rules are turned off, with noreaction of fiscal instruments to deviations of output or real debt fromthe steady state. The only exception is the lump-sum tax, which reactsto deviations of debt (with the debt feedback parameter set close tozero) so as to induce stationarity of the model with a limited fiscalblock. Nevertheless, in the limited fiscal block we allow for rule-of-thumb households, but the estimation results revealed that their sharein this version of themodel is very low. The posterior odds ratio stronglysuggests that themodel containing the full fiscal sector outperforms themodelwith a limited fiscal sector in terms of likelihood (the logmargin-al data density is−2791 for the full model versus−2873 in the modelwith a limited fiscal block).

4. The results

In this section we present the fiscal multiplier values implied by ourstructural DSGEmodel, thenwe use ourmodel to evaluate the impact ofthe selected fiscal measures on the economy.

4.1. Fiscal multipliers

The model's implied fiscal multipliers are listed in Tables 2–4. Thefiscal multipliers are calculated according to Uhlig (2010), so these are


net-present-value multipliers accumulated over time, discounted bythe steady state real interest rate. We list fiscal multipliers with effectson real GDP for individual revenue and expenditure items of the govern-ment budget. The fiscal multipliers are calculated for the case of a tem-porary, one-year fiscal stimulus and for the case of a longer-lasting 10-year fiscal stimulus. The unexpected shocks to the fiscal instrumentsare set so that the ex-anteworsening of the government budget balancein the first year equals 1% of nominal GDP, and the value of the corre-sponding fiscal instrument is kept constant during the affected period.Moreover, the estimated fiscal rule is initially turned off for two years(keeping unaffected fiscal instruments at their steady states), so as toisolate the effects of affected fiscal instruments. Otherwise, keepingthe fiscal rule turned on from the beginning could result in our resultsbeing somewhat blurred by co-movements of fiscal instruments as de-fined in the fiscal rule. We treat the estimated fiscal rule as a good ap-proximation of the fiscal policy settings in the long run, hence wedecided to turn off the fiscal rule at the beginning of the simulations.This also means that the fiscal stimuli in the first two years are fullydebt financed. The role of alternative assumptions, concerning themodel or the simulations, is demonstrated later in the robustness ofthe results. (See Tables 3 and 4.)

Regarding the effect of a temporary fiscal stimulus on real GDP (seeTable 2), the largest effects after the first year occur with governmentconsumption and social security contributions paid by employers,with the fiscal multiplier reaching 0.6. Next, government investmenthas a fiscal multiplier of 0.5, followed by the consumption tax, thewage tax, and unemployment benefits with a corresponding fiscal mul-tiplier of 0.2. The fiscal multipliers for other social benefits, lump-sumtaxes, and the capital tax attain values of roughly 0.1. All the values ofthe fiscal multipliers with effects on real GDP are below 1.

Our values of the fiscalmultipliers for government consumption andinvestment are not far from the CNB estimates of around 0.6 reported inthe study of Hřebíček et al. (2005), which are obtained from empiricalestimates using regression analysis and structural simulation. On the




Table 3Fiscal multipliers, robustness.

Quarters Peak LR

1 4 8 16

Rule-of-thumb households 40%Expenditures (+)



Simple fiscal ruleExpenditures (+)



Elastic labor supply ϕn = 0.5Expenditures (+)


Taxes (−)Consumption tax 0.18 0.25 0.30 0.19 0.30 0.05Wage tax 0.17 0.27 0.42 0.53 0.60 0.59Social contr. employers 0.24 0.70 0.96 0.71 0.96 0.73Capital tax 0.01 0.02 0.05 −0.11 0.05 −0.11Lump-sum tax 0.10 0.08 0.07 −0.09 0.10 −0.14

Note: The ex-ante fiscal stimulus lasts for one year.


other hand, our fiscal multipliers are slightly higher than the ones esti-mated by Klyuev and Snudden (2011) for the Czech Republic using theGIMF model. For instance, the one-year temporary fiscal multiplier forgovernment consumption is larger in our case (0.6) compared to theirestimate (0.4). On the revenue side, our one-year fiscal multipliers forthe consumption tax, the wage tax (0.2), and the capital tax (0.1) areroughly twice as high as the estimates based on the GIMF model. Ac-cording to our model the capital tax has the smallest fiscal multiplier.In fact, this estimate lies within the range found by Coenen et al.(2012b) for the euro area (0.03–0.06).

The fiscalmultipliers for a 10-yearfiscal stimulus have similar valuesin the short run as in the case of a temporary, one-yearfiscal stimulus. Inthe long run, the fiscal multipliers for the 10-year fiscal stimulus aresomewhat lower, and for other social benefits, the consumption tax,the capital tax, and lump-sum taxes the long-run effect on real GDP isslightly negative. Lower fiscal multiplier values for a permanent stimu-lus are confirmed by several other structural models— see Coenen et al.(2012a) for an overview of the effects of fiscal stimuli in the DSGEmodels. The underlying reason is that a longer-lasting stimulus trans-lates into higher government debt, which has to be financed also byhigher taxes. A large increase in taxes leads to a negative wealth effect,which crowds out private demand. Moreover, in our case taxes aredistortionary, which further reinforces the crowding-out effects. Also,the presence of nominal rigidities plays some role, having an effect onthe inflation dynamics (Coenen et al., 2012a, show that the gap betweenshort- and long-run multipliers is higher in models featuring lowernominal rigidities).

The long-run effects of a temporary fiscal stimulus on real GDP (seethe last column in the tables) tell a similar story as in the case of one-year effects, namely, that it is desirable to support the domestic econo-my mainly by increasing government consumption, reducing socialsecurity contributions paid by employers, and fostering government in-vestment. For a longer-lasting fiscal stimulus, the highest effects on realGDP are recorded for taxes associatedwithwages and government con-sumption. Conversely, as regards an appropriate, growth-friendly fiscalconsolidation strategy, hikes in capital or consumption taxes seem de-sirable given the low values of the fiscal multipliers in the long run.These long-run suggestions should be taken with caution, since theyare more or less prone to the specification of the fiscal rule, as will bedemonstrated in the robustness of the results.

4.1.1. RobustnessTo investigate the robustness of the results,we inspect how thefiscal

multipliers changewith: (i) a larger share of rule-of-thumbhouseholds,(ii) a simplified type of fiscal rule, (iii) a different elasticity of laborsupply, (iv) passive monetary policy, and (v) anticipated shocks.

The share of rule-of-thumb households matters for the size of fiscalmultipliers, as can be seen from Table 3. Additionally to the 16% shareof rule-of-thumb households in the baseline, a share of 40% is consid-ered here. Generally, with a higher share of rule-of-thumb householdsthe fiscal multipliers take higher values, in line with evidence fromother literature summarized in the meta analysis of 89 studies byGechert and Will (2012). Lower variation in fiscal multipliers is visiblefor social security contributions paid by employers. This can be ex-plained by the fact that this kind of tax is incurred by firms and doesnot directly affect employees' net labor income. Generally, all the fiscalmultipliers behave according to economic intuition, that is, the higherthe share of rule-of-thumb households in the economy, the strongerthe demand effects generated by consumers in response to fiscalstimuli.

Next, we checked the robustness of the fiscal multipliers with re-spect to the specification of the fiscal rule when the model containsonly a simple calibrated fiscal rule, as adopted from Galí et al. (2007),where lump-sum taxes react to deviations of government consumptionand real bonds from their steady states. With a simple fiscal rule themodel's implied fiscal multipliers in the short run are very similar to


those obtained with a more comprehensive estimated fiscal rule. Thisis due to the fact that the fiscal rule is switched off for the initial twoyears by assumption. So, again, the highest temporary one-year fiscalmultipliers are found for government consumption, social security con-tributions paid by employers, and government investment. As to the cu-mulative long-run gains in the case of a one-year fiscal stimulus, thesituation changes. Here, the largest effects are attained for lump-sumtaxes, social security contributions paid by employers, and governmentconsumption. The relevance of lump-sum taxes in this case is the resultof the fiscal rule used, as only lump-sum taxes can adjust in response torising government debt and government consumption, in contrast tothe comprehensive estimated fiscal rule, where all fiscal instrumentscan adjust to variations in government debt and intermediate production.

Originally, in the g3 model the utility function is linear in labor sup-ply, implying an infinitely elastic supply of labor. By contrast, in ourmodel we estimated the inverse of the Frisch elasticity of labor supplyϕn at a value of roughly 3. Therefore, we check the sensitivity of this pa-rameter on the values of the fiscal multipliers. Alternatively, we set thevalue of parameterϕn=0.5. On inspecting the results in Table 3, we seethat thefiscalmultipliers have increased. This is due to a greaterwilling-ness of households to supply labor to firms, which want to cover thehigher demand generated by the fiscal stimuli. The ordering of thefirst three fiscal instruments, in terms of GDP growth contribution, haschanged slightly. In the short run it is the same: the most significantare government consumption, social security contributions paid by em-ployers, and government investment, but in the long run social security




Table 4Fiscal multipliers, robustness.

Quarters Peak LR

1 4 8 16

Passive monetary policyExpenditures (+)



Anticipated stimulusExpenditures (+)


Taxes (−)Consumption tax 0.18 0.21 0.23 0.10 0.23 −0.03Wage tax 0.16 0.21 0.30 0.28 0.30 0.26Social contr. employers 0.40 0.82 0.83 0.44 0.88 0.42Capital tax 0.02 0.03 0.06 −0.06 0.06 −0.01Lump-sum tax 0.10 0.08 0.08 −0.04 0.10 −0.07

Table 5Selected fiscal consolidation measures (in % of GDP).Source: Ministry of Finance of the Czech Republic and CNB estimates.

Consolidation measure 2013 2014 2015 Model'svariable

Direct taxes7% additional personal income tax 0.0 0.1 0.1 τW

Abolition of pensioners' allowance 0.1 0.1 0.1 τW

Lower deductions for self-employed 0.1 0.1 0.1 τW

Withholding tax against tax havens, abolitionof health insurance ceiling

0.1 0.1 0.1 τW, τS

Indirect taxes1 pp increase in VAT 0.5 0.5 0.4 τC

Abolition of excise exemptions 0.0 0.1 0.1 τC

Other taxes1 pp increase in real estate tax 0.1 0.1 0.1 τK

Sales of emission permits 0.1 0.1 0.1 τK

ExpenditureFreezing wages for state employees −0.1 −0.3 −0.4 GSavings in state administration 0.0 −0.3 −0.6 GLower subsidies −0.1 −0.1 −0.1 Ig

Lower social benefits/pensions −0.3 −0.4 −0.5 OBTotal measures 1.4 2.0 2.6

Note: Impacts of consolidation measures accumulate over time.


contributions paid by employers are followed by government consump-tion and the wage tax.

Furthermore, the fiscal multipliers change when monetary policy ispassive, which means that monetary policy does not follow the Taylorrule, i.e., the central bank does not raise interest rates in response to ris-ing inflation. We simulate the effects of fiscal stimuli with temporarilypassive monetary policy, when the interest rate is kept constant fortwo years at its steady state level and thereafter adjusts according tothe model's Taylor rule. This simulation is different from the situationwhere interest rates hit the zero lower bound; nonetheless, the direc-tion of real effects might be similar. The fiscal multipliers with passivemonetary policy are reported in Table 4. Many of the fiscal multipliershave increased, with the highest one-year multipliers being attainedby government consumption, government investment, and the wagetax. The increase in the fiscal multipliers when monetary policy is pas-sive is due to the fact that the monetary conditions are looser in thefirst two years compared to the baseline. The results for social securitycontributions paid by employers are puzzling, as the fiscal multipliershave decreased compared to the baseline.

As a last robustness check, fiscal multipliers with fully anticipatedshocks to fiscal instruments were calculated. The majority of the fiscalmultipliers have decreased or stayed roughly the same compared tothe baseline. One notable exception concerns social security contribu-tions paid by employers in the short run. The highest one-year fiscalmultipliers are recorded by social security contributions paid by em-ployers (0.8), government consumption (0.6), and government invest-ment (0.4). The lower real GDP effects might be explained by the factthat households fully expect the tax cuts or expenditure increases tobe temporary and choose rather to smooth their consumption profiles.Another explanation might be that all the fiscal instruments have sub-stantial backward-looking elements by definition.

4.2. The impact of selected fiscal measures

Having built and estimated our model, we try to illustrate its usewith a real-world example from the Czech economy. In April 2012,the Czech government announced a new fiscal consolidation packageof various measures for 2013–2015. We focus on selected, already


approved measures from this consolidation package, with further de-tails provided in Table 5. Obviously, it is interesting to quantify themac-roeconomic effects of such consolidation measures on the Czecheconomy.

Firstly, in order to simulate the macroeconomic effects of the select-ed fiscal measures, we mapped these measures onto the models' fiscalinstruments. Afterwards, the estimated impacts of the individual mea-sures on the government budget balance, expressed in percent of nom-inal GDP, were transformed into desired deviations of the relevant fiscalrevenues/expenditures using the steady state value of themodel's nom-inal GDP. Then the paths of the affected fiscal revenues/expenditures forthe calculated deviations are adjusted in the conditional simulation. Tobe more specific, unexpected shocks to fiscal revenues/expendituresare endogenously set so as to deliver the desired exogenous paths ofthe affected fiscal revenues/expenditures. The remaining unaffected fis-cal instruments (unemployment benefits and lump-sum transfers)were kept fixed at their steady state levels for two years; thereafter,the estimatedfiscal rule is switched on, serving as a good approximationof the fiscal policy settings in the long run.

The results given by our model show that the selected consolidationmeasures might imply a tightening of real GDP growth by 0.4, 0.8, and1.1 percentage points in the respective years compared to the baselinescenario with unchanged fiscal policy. Given the size of the consolida-tion measures selected, the implicit fiscal multiplier of the simulatedmeasures is roughly 0.4. Tomake our simulationmore realistic, we arbi-trarily increase the share of “rule-of-thumb” households to 40%, so as toreflect the idea that during downturns there are more households whoare unable to save. In this case, the selected consolidation measureswould depress real GDP growth by 0.7, 1.2, and 1.7 percentage points,with the average implicit fiscal multiplier increasing to 0.6. Moreover,to mimic the current state of the economywhen interest rates have vir-tually hit the zero lower bound, in another simulation we set monetarypolicy as passive for three years (e.g. the central bank does not react todeviations in inflation). In this setting, the impacts on real GDP growthwould increase to 0.7, 1.4, and 2.0 percentage points, associated withan average implicit multiplier of 0.7. The calculated impacts should betaken with caution, since there is a further simplification stemmingfrom the fact that the fiscal shocks in our model are treated as tempo-rary (i.e., as having only temporary effects which do not affect thesteady state of themodel), whereas some of the above-mentioned fiscal





measures might be valid permanently, thus pushing the economytoward a new equilibrium.

5. Conclusion

In this paper we presented a satellite DSGE model for the CzechRepublic to study the effects of fiscal policy on the economy. OurDSGE model shares key features of the CNB's core g3 model, developedby Andrle et al. (2009), and contains a comprehensive fiscal block. Themost distinctive fiscal features of our DSGE model are due to theinclusion of “rule-of-thumb” consumers, detailed specification of therevenue and expenditure fiscal instruments, productive governmentspending, productive government capital, unemployment dynamics,and the estimated fiscal rule.

The real GDPfiscalmultipliers fromour DSGEmodel suggest that thelargest multipliers after the first year of a temporary fiscal stimulus areassociated with government consumption and social security contribu-tions paid by employers (both 0.6) and government investment (0.5),followed by the consumption tax, the wage tax and unemploymentbenefits (all roughly 0.2), then by other social benefits, lump-sumtaxes, and the capital tax (0.1). Thus, our results imply that the strongesteffects, in terms of real GDP, are associated with an increase in govern-ment consumption and a decrease in social security contributions paidby employers and an increase in government investment; or, to put itdifferently, fiscal consolidations based on cuts in government consump-tion, increases in social security contributions paid by employers or cutsin government investment would be associated with substantial realGDP losses, at least in the short to medium run. Conversely, a less costlyway to consolidate public finance is to raise capital taxes, because oftheir low impact on GDP growth.

For practical purposes, we used our model to perform a partial eval-uation of the effects on the real economy of selected consolidationmea-sures related to the ongoing process of fiscal consolidation in the CzechRepublic. Conditional on the model's set of simplifying assumptions(inter alia, limited supply-side aspects and the assumption of temporaryfiscal shocks) we found considerable impacts on real GDP dynamics,specifically impacts of 0.4, 0.8, and 1.1 percentage points in 2013,2014, and 2015, as compared to the baseline with unchanged fiscalpolicy. These further increase if a higher share of “rule-of-thumb”households and/or passive monetary policy is considered.

This paper could be extended in several directions. The robustness ofour results could be further checked in terms of the underlying modelmechanisms and assumptions. For example, what is the role of comple-mentarity/substitutability between private and government consump-tion/capital in the measured values of fiscal multipliers? And what isthe size of the fiscal multipliers if the zero lower bound on nominal in-terest rates is reached?One could also further refine thefiscal sector; forinstance, it is possible to further elaborate government labor servicesand model them explicitly as a production input. Another possible ex-tension would be to build a DSGE-VAR model, that is, to simulate thepriors from our DSGE model and use them afterwards in the estimationof the Bayesian VAR model.

Acknowledgments

This research was supported by the Czech National Bank ResearchProject No. D2/10. We are grateful to Michal Franta, Philipp Hartmann,


Josef Hollmayr, Michal Kejak, Marco Ratto, Sergey Slobodyan, Jan in'tVeld, andMilan Výškrabka for helpful discussions, and to the two anon-ymous referees for their valuable suggestions. The paper benefited fromcomments at CNB seminars, the UECE Conference on Economic andFinancial Adjustments in Europe, Lisbon, 2013, and the IWH and KielEconomics Workshop on Fiscal Policy and the Great Recession, Halle(Saale), 2013. The views expressed in this article are those of theauthors and do not necessarily reflect the views of their institutions.

References

Ambriško, R.,Babecký, J.,Ryšánek, J.,Valenta, V., 2012. Assessing the impact of fiscal mea-sures on the Czech economy. Czech National Bank Working Paper, No. 15/2012.

Andrle, M.,Hlédik, T.,Kameník, O.,Vlček, J., 2009. Implementing the new structural modelof the Czech National Bank. Czech National Bank Working Paper, No. 2/2009.

Bailey, M.J., 1971. National Income and the Price Level: A Study inMacroeconomic Theory,2nd edition. McGraw-Hill, New York.

Barrell, R.,Holland, D.,Pain, N.,Kovacs, M.A.,Jakab, Z.,Smidkova, K.,Sepp, U.,Cufer, U., 2004.An econometric macro-model of transition: policy choices in the pre-accessionperiod. Macroeconomics 0403004 (EconWPA).

Barro, R.J., 1981. Output effects of government purchases. J. Polit. Econ. 89 (6),1086–1121.

Baxter, M.,King, R.G., 1993. Fiscal policy in general equilibrium. Am. Econ. Rev. 83 (3),315–334.

Christoffel, K.,Coenen, G.,Warne, A., 2008. The new area-wide model of the euro area — amicro-founded open-economy model for forecasting and policy analysis. WorkingPaper Series, 944. European Central Bank.

Coenen, G., Straub, R., Trabandt, M., 2011. Fiscal policy and the great recession in the Euroarea. mimeo.

Coenen, G.,Erceg, C.J.,Freedman, C.,Furceri, D.,Kumhof, M.,Lalonde, R.,Laxton, D.,Linde, J.,Mourougane, A.,Muir, D.,Mursula, S.,de Resende, C.,Roberts, J.,Roeger,W.,Snudden, S.,Trabandt, M., in 't Veld, J., 2012a. Effects of fiscal stimulus in structural models. Am.Econ. J. Macroecon. 4 (1), 22–68.

Coenen, G.,Straub, R.,Trabandt, M., 2012b. Fiscal policy and the great recession in the euroarea. Am. Econ. Rev. 102 (3), 71–76.

Franta, M., 2012. The effects of fiscal policy in the Czech Republic: evidence based on var-ious identification approaches in a VAR framework. Czech National Bank WorkingPaper, No. 13/2012.

Galí, J., 2011. The return of the wage Phillips curve. J. Eur. Econ. Assoc. 9 (3), 436–461.Galí, J.,López-Salido, J.D.,Vallés, J., 2007. Understanding the effects of government spend-

ing on consumption. J. Eur. Econ. Assoc. 5 (1), 227–270.Gechert, S.,Will, H., 2012. Fiscal multipliers: a meta regression analysis. Macroeconomic

Policy Institute Working Paper, 97.Hřebíček, H., Král, P., Říkovský, M., 2005. An up-date of fiscal impulse quantification,

(Aktualizace propočtu fiskálního impulsu, in Czech only), Czech National Bank,mimeo.

Ipsos Tambor, 2012. 40 procent Čechů bude žít od výplaty k výplatě (in Czech only)(January).

Klyuev, V.,Snudden, S., 2011. Effects of fiscal consolidation in the Czech Republic. Czech J.Econ. Financ (Finance a uver) 61 (4), 306–326.

Kumhof, M.,Muir, D.,Mursula, S., Laxton, D., 2010. The global integrated monetary andfiscal model (GIMF) — theoretical structure. International Monetary Fund WorkingPaper, 10/34.

Leeper, E.M.,Plante, M.,Traum, N., 2010. Dynamics of fiscal financing in the United States.J. Econ. 156 (2), 304–321.

Prušvic, D., 2010. Fiscal stimuli and fiscal multipliers in small open economies: the case ofthe Czech Republic. Presented on the 40th Conference on Medium Term EconomicAssessment, September 23–25, 2010, Luxembourg.

Ratto, M.,Roeger, W., in 't Veld, J., 2009. Quest III: an estimated open-economy DSGEmodel of the euro area with fiscal and monetary policy. Econ. Model. 26 (1),222–233.

Štork, Z.,Závacká, J., 2010. Macroeconomic implications of fiscal policy measures in DSGE.Ministry of Finance of the Czech Republic Working Paper, No. 1/2010.

Uhlig, H., 2010. Some fiscal calculus. Am. Econ. Rev. 100 (2), 30–34.Valenta, V., 2011. Interactions Between Fiscal Policy and Real Economy in the Czech

Republic: A Quantitative Analysis(Doctoral Dissertation) University of Economics,Prague.


http://refhub.elsevier.com/S0264-9993(14)00276-4/rf0075


















































assessing the impact of fiscal measures on the czech economy

Documents