Ma a in, Luigi; Salo i, Simone
Wo king Pape
Consump ion Mul iplie s o Di e en Types o Public
Spending: a S uc u al Vec o E o Co ec ion Analysis o
he UK
Quade ni - Wo king Pape DSE, No. 719
P o ided in Coope a ion wi h:
Uni e si y o Bologna, Depa men o Economics
Sugges ed Ci a ion: Ma a in, Luigi; Salo i, Simone (2010) : Consump ion Mul iplie s o Di e en
Types o Public Spending: a S uc u al Vec o E o Co ec ion Analysis o he UK, Quade ni -
Wo king Pape DSE, No. 719, Alma Ma e S udio um - Uni e si à di Bologna, Dipa imen o di
Scienze Economiche (DSE), Bologna,
h ps://doi.o g/10.6092/unibo/amsac a/4522
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/159560
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by-nc/3.0/
Consump ion mul iplie s o di e en
ypes o public spending:
a s uc u al ec o e o co ec ion
analysis o he UK
Luigi Ma a in
Simone Salo i
Quade ni - Wo king Pape DSE N° 719
Consump ion mul iplie s o di e en ypes o public
spending: a s uc u al ec o e o co ec ion analysis o
he UK
Luigi Ma a in
Depa men o Economics, Uni e si y o Bologna, S ada Maggio e 45, 40125 Bologna (I aly)
+39 051 2092606
luigi.ma a [email protected]
Simone Salo i
Depa men o Economics, Na ional Uni e si y o I eland, Galway, Uni e si y Road, Galway (I eland)
+353 091 493053
simone.salo [email protected]
We hank all he pa icipan s o he 2011 Annual Con e ence o he Sco ish Economic Socie y o use ul
commen s. A p e ious e sion o his pape was ci cula ed unde he i le “The esponse o p i a e consump ion
o di e en public spending ca ego ies: VAR e idence o he UK”: we hank all he pa icipan s o he Ecomod
2010 con e ence and o he Depa men o Economics semina in Bologna in No embe 2009 o use ul
commen s on ha e sion. All emaining e o s a e ou esponsibili y.
This e sion: 14/05/2011
Abs ac . The aim o his pape is o in es iga e he ela ionship be ween go e nmen spending and p i a e consump ion in
he UK, o which he e is sca ce p e ious empi ical e idence. We disagg ega e public expendi u e in o h ee ca ego ies and
sea ch o he co esponding p i a e consump ion mul iplie s. Ou analysis includes he es ima ion o a s uc u al ec o e o
co ec ion model, using qua e ly non-in e pola ed da a o he pe iod 1981:1 – 2007:4. Ini ially, we es ima e nega i e (and
ba ely signi ican ) e ec s on consump ion o shocks o o al public spending. Then, using he spending b eaking down, we
ind ha while shocks o public wages c owd-ou p i a e consump ion as p edic ed by neoclassical models, shocks o he
non-sys ema ic componen o social spending and go e nmen pu chases o goods and se ices gene a e a posi i e eac ion,
so o c owd-in p i a e consump ion. Thus, he quali a i e and quan i a i e dimensions o iscal mul iplie s on p i a e
consump ion change ac oss di e en public spending ca ego ies. Ou indings sugges ha any empi ical suppo o
compe ing heo e ical models on he issue would bene i om a disagg ega ion o go e nmen expendi u e, a he han
ocusing on he agg ega e measu e.
Keywo ds: Fiscal policy, Coin eg a ion, S uc u al VECM, Subse SVECM.
JEL classi ica ion: E62, H30.
2
1. In oduc ion
A he beginning o 2009, when he need o implemen an unp eceden ed inc ease in public spending o
coun e ac he global ecession became clea , a common ques ion among policy make s and T easu y
Depa men s all o e he wo ld was he ollowing: which pa o public expendi u e should we p i ilege
in o de o maximize he posi i e e ec on p i a e demand? Two yea s la e , as he agenda is domina ed
by he need o iscal e enchmen , he ques ion is e e sed: wha should be cu i s in o de no o
dampen he eco e y? In bo h cases, he issue o be add essed is he same: inding he public expendi u e
ca ego y associa ed wi h he highes consump ion mul iplie . A emp s o e alua e he e ec i eness o
iscal policy o en ely on he app aisal o he sign and magni ude o public expendi u e’s impac ei he
on GDP (e.g. Blancha d and Pe o i 2002), o p i a e consump ion (e.g. Ga cía and Ramajo 2005), o on
p i a e in es men (e.g. Wang 2005). In ou pape we in es iga e he p i a e consump ion mul iplie
associa ed no only wi h agg ega e go e nmen spending, bu also wi h each o i s h ee di e en
componen s, namely wage, non-wage and social secu i y expendi u es (see Figu e 1).
FIGURE 1 ABOUT HERE
These h ee spending ca ego ies di e in many espec s, so i is easonably plausible o hypo hesize
di e en quali a i e and quan i a i e impac s on p i a e consump ion. The undamen al mo i a ion o
ou wo k is he belie ha a ocus on he agg ega e measu e o public spending, a he han on a mo e
accu a e disagg ega ion, migh imply inaccu a e policy conclusions on i s s imula ing ole (i any).
Pa icula ly, we explici ly conside go e nmen social expendi u e as a spending ca ego y pe se, a he
han including i in o he e enue componen as usually done in he li e a u e. The a ionale o ha
elies in ou desi e o e i y he consump ion-enhancing p ope ies o income edis ibu ion, which is
implemen ed h ough he ne social secu i y ans e s.
Ou empi ical analysis is based on a s uc u al ec o e o co ec ion (SVEC) model using UK qua e ly
non-in e pola ed da a (1981:Q1 -2007:Q4). A he momen he e idence on his coun y is sca ce, as
only ew s udies ha e employed UK da a (Pe o i 2004 and 2007, Ramos and Roca-Sagales 2008,
Monacelli and Pe o i 2010). Mo eo e – as all o hem adop ed VAR es ima ions - none o hem
add essed he coin eg a ing p ope ies o he se ies unde in es iga ion. In pa icula , o he pe iod
1981-2001, Monacelli and Pe o i (2010) ind mixed esul s on he e ec s o go e nmen spending on
p i a e consump ion, wi h posi i e e ec s when using he GDP de la o and nega i e e ec s when using
own de la o s.1 In es iga ing a simila pe iod, Pe o i (2004) inds ha he cumula ed esponse o
consump ion o a 1% o GDP go e nmen spending shock a e h ee yea s is me ely +.05% o GDP.
Finally, wi h da a up o 2005, Ramos and Roca-Sagales (2008) ind ha go e nmen spending
nega i ely a ec s GDP.2 Ou aim is o shed ligh on hese ew (and mixed) esul s analyzing UK da a
using a me hodology which inno a es in se e al aspec s. Fi s o all, we in es iga e public spending
using bo h i s agg ega e measu e and each o i s componen s. This ma ks a di e ence wi h wha has
been done in he li e a u e so a (Bee sma e al. 2006 and 2008, Gio dano e al. 2007, Ca allo 2005 and
2007), as we explici ly conside ne social secu i y spending in he analysis ins ead o including i in o
he ne axes a iable. Then, we use a coin eg a ing amewo k ha pe mi s o ex ac in o ma ion om
he se ies unde in es iga ion aking in o accoun hei long un p ope ies, in a spi i close o K usec
(2003) who es ima ed a SVEC model o iscal policy using US da a. Di e en ly om him, we include
public deb in he model, ollowing Fa e o and Gia azzi (2007).
Ou esul s show ha shocks o agg ega e public spending nega i ely a ec p i a e consump ion, e en i
he es ima es a e ba ely s a is ically di e en om ze o a s anda d le els. Looking a he e ec s o he
h ee componen s o public spending, we ind ha shocks o bo h ne social secu i y ans e s and
go e nmen “pu e” consump ion (pu chases o goods and se ices) esul in posi i e esponses o
p i a e consump ion. While he esponse o he i s shock las s ou yea s, he esponse o he second
shock dies ou wi hin six qua e s. On he o he hand, shocks o public wages nega i ely a ec p i a e
consump ion, wi h e ec s ha die ou less han wo yea s a e he shock. These esul s suppo he
in ui ion ha using o al go e nmen spending does no seem o be a easonable simpli ica ion when
s udying he e ec s o public spending on p i a e consump ion.
We belie e hese esul s o be ele an o he heo e ical deba e be ween al e na i e and compe ing
app oaches modelling p i a e consump ion’s impac o iscal shocks. In pa icula , he s anda d
neoclassical eal business cycle (RBC) model p edic s a all in consump ion ollowing a go e nmen
1 In ac , he consump ion impulse esponses show ha , e en when using he GDP de la o , he cumula ed esponse o p i a e
consump ion is ci ca ze o 16 qua e s a e he public spending shock ( en qua e s a e he shock he e ec s become
nega i e, simila ly o wha Pe o i 2007 inds in es iga ing a la ge ime span).
2 The e a e o he s udies o he UK iscal policy (e.g. Sawye 2007, Budd 2010), bu hei scopes a e di e en om he one o
quan i ying he e ec s o public spending on GDP o on some componen s o agg ega e demand.

4
spending shock, because o he Rica dian equi alence: highe public spending mus be ma ched by an
equi alen inc ease in axa ion in p esen discoun ed e ms. Thus, in e empo al op imizing consume s
su e om a nega i e weal h e ec ha dec eases consump ion (wi h posi i e e ec s on ou pu due o
he inc eased labou supply igge ed by he weal h e ec ). Since i ually no s udy seems o con i m he
p edic ion o he s anda d neoclassical model (as poin ed ou by Galí e al. 2007), New Keynesian
economis s a emp ed o econcile heo y wi h empi ical e idence by escuing a consump ion-enhancing
ole o iscal policy. This has been accomplished ei he using ini e-ho izons amewo ks (Blancha d
1985) o in oducing c edi -cons ained agen s and ule-o - humb consume s (Mankiw 2000, Galí e al.
2004 and 2007, Coenen and S aub 2005).3 Abo e all, his la e app oach has gained conside able
a en ion. I assumes ha he e is a ac ion o non-Rica dian households who do no op imize o e he
li e cycle and a e hus o ced o consume ou o cu en income, so ha hei consump ion esponds
p omp ly o a iscal policy impulse.4 A u he esea ch s and explici ly conside s he pe se go e nmen
expendi u e’s impac on consump ion. This is o en ca ied ou by an ad-hoc u ili y unc ion
speci ica ion whe e p i a e and public consump ion a e en e ed in a non-addi i e o m, so o ob ain a
non-ze o impac o one on he ma ginal u ili y o he o he (Bouakez and Rebei 2003).
Ou s udy can be ele an o all he abo e-men ioned heo e ical discussions. We p o ide e idence ha
conside ing agg ega e go e nmen spending can indeed be misleading. The iden i ica ion o social
expendi u e and pu e consump ion as he sole expendi u e ca ego ies which a e e ec i e in s imula ing
p i a e consump ion leads o wo ema ks. Fi s , he complemen a i y/subs i u abili y issue canno be
discussed independen ly om a su icien disagg ega ion o go e nmen spending. Second, he ule-o
humb-consume s app oach can indeed be jus i ied no longe on he assump ion o an exogenous
ac ion o c edi cons ained agen s. Fo ins ance, i could be hypo hesized ha he consump ion o he
bene icia ies o social secu i y ans e s (p esumably lying in he lowe pa o income dis ibu ion) is
3 As a ma e o ac , he e a e also al e na i e ways o he same esul . Ra n e al. (2004) ob ain a posi i e e ec on
consump ion wi hou c edi -cons ained agen s, bu assuming ha he ep esen a i e indi idual o ms consump ion habi s on
he indi idual a ie y in a monopolis ic compe i ion se ing, a he han on agg ega e consump ion. Co se i e al. (2009) use
spending e e sals o ge a posi i e esponse o p i a e consump ion o go e nmen spending shocks.
4 As discussed by Galí e al. (2007), he p esence o non-Rica dian consume s mus be coupled wi h s icky p ices and
impe ec ly compe i i e ma ke s in o de o ob ain a p i a e consump ion’s posi i e esponse.
s imula ed by a leas one componen o public spending, hus he ule-o - humb models could be
pa ame e ized acco dingly.
The emainde o his pape is o ganized as ollows. Sec ion 2 p esen s he heo y behind he SVEC
me hodology. Sec ion 3 con ains he es ima es o he VEC model. Sec ion 4 deals wi h he s uc u al
analysis (iden i ica ion assump ions and impulse esponse analysis). Sec ion 5 concludes and discusses
some policy implica ions.
2. The SVEC me hodology
The li e a u e ha in es iga es he e ec s o public spending on p i a e consump ion and ou pu has
la gely used s uc u al VAR models since he seminal a icle by Blancha d and Pe o i (2002) on he US
economy. In ou pape we use a closely ela ed echnique: a SVEC model ha akes in o accoun he
coin eg a ing p ope ies o he se ies unde in es iga ion. The analysis s a s om a educed o m VEC
model:
'1 1 1 1 1
... ,
p p
y y y y D u

    
        
(2.1)
whe e y is a Kx1 ec o o ime se ies, D is a ec o o de e minis ic e ms, and Γ1, ..., Γp-1 a e KxK
coe icien ma ices. The educed o m dis u bance ( o ecas e o ) u is a Kx1 unobse able ze o mean
whi e noise p ocess wi h co a iance ma ix Σu. α and β a e Kx ma ices con aining he loading
coe icien s and he coin eg a ing ec o s, espec i ely. We a e in e es ed in he e ec s o undamen al
shocks ε (e o e ms o he s uc u al o m VEC model) on he sys em a iables y .. The link be ween
he wo e o s is gi en by he con empo aneous impac ma ix A:
uA


(2.2)
In o de o analyze he e ec s o he unde lying s uc u al shocks we need o iden i y he A ma ix, by
eco e ing i s K2 elemen s. Assuming ha he s uc u al shocks a e unco ela ed and ha e uni a iances
(Σε=Ik) we ge :
' ' ' '
u
E uu E A A AA


   
     
   
(2.3)
The symme y o Σε and he no maliza ion o he s uc u al a iances impose K(K+1)/2 es ic ions on
he K2 pa ame e s o A. The e o e, o exac ly iden i y he elemen s o A, we need o impose K(K-1)/2
6
addi ional linea ly independen es ic ions. We p e e o impose long un es ic ions a he han
con empo aneous es ic ions, which a e usually ha de o jus i y e en wi h qua e ly da a. F om
G ange 's ep esen a ion heo em (Johansen, 1995) i ollows ha he long un e ec s o s uc u al
shocks ε can be w i en as:
 
1,CA
(2.4)
whe e C(1) is he o al impac ma ix compu ed as:
 
 
 
1
' 1 '
1
1p
K i i
CI
   



      
(2.5)
β┴ and α┴ ep esen he o hogonal complemen s o β and α espec i ely. No e ha C(1) has educed
ank k(C(1))=K- . Long un es ic ions can now be imposed by se ing elemen s o (2.4). As in he
common ends li e a u e (e.g. see King e al. 1991), we dis inguish be ween pe manen and ansi o y
e ec s. Se ing elemen s o (2.4) o ze o and eplacing C(1) by an es ima o ob ained om he educed
o m, he es ic ions may be w i en in implici o m as:
   
   
*
ˆˆ
(1) (1) 0,
l l k l
R ec C A R I xC ec A R ec A  
(2.6)
whe e Rl is an app op ia e es ic ion ma ix. These long un es ic ions can be combined wi h
con empo aneous es ic ions on he elemen s o A. Wi h enough iden i ying es ic ions, es ima es o
he con empo aneous impac ma ix can be ound by maximizing he concen a ed log-likelihood
unc ion wi h espec o he ee s uc u al pa ame e s (see e.g. B ei ung e al. 2004). Concen a ing he
log-likelihood wi h espec o he educed o m pa ame e s is no longe possible i addi ional es ic ions
o α, Γ1, ..., Γp-1 a e imposed. Ne e heless, esiduals om a subse VEC model may s ill gi e a
easonable es ima e o Σu (Hamil on, 1994). Thus, he same es ima ion echnique as be o e can be used
o o m easonable es ima es o A.
Based on his discussion, he econome ic analysis o he UK iscal policy e ec s will be he ollowing:
i s , we de e mine he coin eg a ion ank o he sys em o in e es , imposing o e -iden i ying
es ic ions on he coin eg a ing ec o s using he Johansen (1995) me hod. These coin eg a ing ela ions
a e used o se up a ull VEC model, whe e no u he es ic ions a e imposed. Residuals om he ull
and subse VEC model a e used o o m es ima es o Σu. Second, long un and con empo aneous
es ic ions a e used o o m es ima es o A. Using he es ima ed con empo aneous impac ma ix, we
eco e he s uc u al shocks and we analyze hei impac h ough an impulse esponse analysis.
3. Analysis o UK public expendi u e
The p esen sec ion opens ou empi ical analysis on he UK economy. Pa icula ly, i is conce ned wi h
he desc ip ion and discussion o he da ase (subsec ion 3.1), he in es iga ion o he coin eg a ing
p ope ies (subsec ion 3.2), and he es ima ion o he VEC model (subsec ion 3.3).
3.1 Da a desc ip ion
“The a ailabili y o qua e ly iscal a iables ep esen s he main cons ain o he analysis o iscal
policy wi h VAR models” (Gio dano e al. 2007, p. 710). Fu he mo e, Pe o i (2004) co ec ly wa ns
agains he dis o ions coming om he usage o qua e ly da a se ob ained by in e pola ion o yea ly
alues. This ema k makes he da a a ailabili y cons ain e en mo e binding, and poses conside able
limi a ions o he implemen a ion o a ully-equipped la ge scale ime se ies analysis. We ha e chosen o
sac i ice he gene ali y o ou empi ical esul s in a ou o a comple e non-in e pola ed qua e ly
da ase . We use UK da a om he i s qua e o 1981 o he las qua e o 2007 aken om he OECD
Economic Ou look no. 83.5 We use he ollowing a iables: he log o eal go e nmen wage
expendi u e GW , he log o eal go e nmen ne social secu i y spending (cyclically adjus ed) GSS , he
log o eal go e nmen spending in goods and se ices GC , he log o eal go e nmen consump ion
(calcula ed as he sum o he h ee p e ious a iables), he log o eal o al go e nmen consump ion,
GTOT , he log o eal axes (cyclically adjus ed) T , he log o eal p i a e consump ion C , and he log
o eal go e nmen ne inancial liabili ies B . The ime se ies a e shown in Figu e 2. The de ails on he
cons uc ion o he iscal a iables used in he sys em and some desc ip i e s a is ics a e gi en in he
Appendix. We do no include public in es men in o he analysis as his componen implies ex e nal
e ec s, such as p oduc ion ex e nali ies, ha a e no immedia ely associa ed wi h p i a e consump ion,
which is he ocus o his pape (see e.g. Mi nik and Neumann 2001). We also do no explici ly conside
5 The qua e ly da a o he Economic Ou look a e no mally ob ained by in e pola ion, bu no hose o he UK. The beginning
o he pe iod has been chosen because o he s ong e idence o a s uc u al b eak in UK da a be ween 1981 and he p e ious
pe iod (Pe o i, 2004).
14
an empi ical analysis ocused on he UK economy, as p e ious li e a u e on his coun y con ains sca ce
and ambiguous e idence. We employ a SVEC model wi h an iden i ica ion s a egy ha akes ad an age
o he di e en ia ion be ween pe manen and empo a y shocks in he coin eg a ing amewo k. Ou
esul s con o m o he p e ious e idence when using public spending in i s agg ega e, as we es ima e
nega i e e ec s on p i a e consump ion which a e ba ely di e en om ze o a s anda d con idence
le els (Pe o i 2004 and 2007, Ramos and Roca-Sagales 2008, Monacelli and Pe o i 2010). In e es ing
insigh s come om he es ima ion o he model using he h ee componen s o public spending, since we
ind ha shocks o bo h ne social secu i y ans e s and go e nmen pu e consump ion esul in posi i e
esponses o p i a e consump ion. While he esponse o he i s one las s o a long pe iod o ime wi h
signi ican e ec s up o ou yea s om he shock, he esponse o he pu e consump ion shock dies ou
wi hin six qua e s. This consump ion-enhancing ole o iscal policy is en isaged in New Keynesian
models. On he o he hand, shocks o public wages nega i ely a ec p i a e consump ion (as p edic ed
by s anda d RBC models), wi h e ec s ha die ou wi hin wo yea s a e he shock. These esul s
suppo he in ui ion ha using o al agg ega e go e nmen expendi u e does no seem o be a easonable
simpli ica ion when s udying he e ec s o public spending on p i a e consump ion. While we belie e
ha his analysis can ep esen a use ul con ibu ion o a mo e e ec i e managemen o iscal policy
ools on he expendi u e side, he gene al alidi y o he indings is ce ainly limi ed by he ocus on a
single coun y. A panel analysis on Eu opean Union coun ies would allow a mo e comple e answe o
ou o iginal ques ion and would p obably be he bes nex s ep o be aken.
Re e ences
Alesina AF, A dagna S (2009) La ge changes in iscal policy: axes e sus spending. NBER wp
Bee sma RMWJ, Bo enbe g AL (1998) Mone a y union wi hou iscal coo dina ion may discipline policymake s. J In Econ
45:239-258
Bee sma RMWJ, Giuliodo i M, Klaassen F (2006) T ade spill-o e s o iscal policy in he Eu opean Union: a panel analysis.
Econ Pol 21(48):639-687
Bee sma RMWJ, Klaassen F, Giuliodo i M (2008) The e ec s o public spending shocks on ade balances and budge
de ici s in he Eu opean Union. JEEA 6(2-3):414-423
Blancha d O (1985) Deb , de ici s and ini e ho izons. J Poli Economy 93(2):223-247

Blancha d O, Pe o i R (2002) An empi ical cha ac e isa ion o he dynamic e ec s o changes in go e nmen spending and
axes on ou pu . Qua J Econ117(4):1329-1368
Bouakez H, Rebei N (2003) Why does p i a e consump ion ises a e a go e nmen spending shock? Bank o Canada wp
B ei ung J, B üggemann R, Lü kepohl H (2004) S uc u al ec o au o eg essi e modeling and impulse esponses. In:
Lü kepohl H, K ä zig M (ed) Applied ime se ies econome ics. Camb idge Uni e si y P ess, pp 159-196
B üggemann R, Lü kepohl H (2001) Lag selec ion in subse VAR models wi h an applica ion o a U.S. mone a y sys em. In:
F ohn J, F iedman R (ed) Econome ic s udies: a es sch i in honou o Joachim F ohn. LIT Ve lag Müns e
Budd A (2010) Fiscal policy unde labou . Na Ins Econ Re 212:R34-R48
Ca allo M (2005) Go e nmen employmen expendi u e and he e ec s o iscal policy shocks. Fed Res Bank o San
F ancisco
Ca allo M (2007) Go e nmen consump ion expendi u e and he cu en accoun . PFM 7:73-115
Coenen G, S aub R (2005) Does go e nmen spending c owd in p i a e consump ion? Theo y and empi ical e idence o he
Eu o a ea. In Finance 8(3):435-470
Co se i G, Meie A, Mülle G (2009) Fiscal s imulus wi h spending e e sals. IMF wp 09/106
Da idson R, MacKinnon J (1993) Es ima ion and in e ence in econome ics. Ox o d Uni e si y P ess, London
Fa e o CA, Gia azzi F (2007) Deb and he e ec s o iscal policy. NBER wp
Galí J, López-Salido JD, Vallés J (2004) Rule-o - humb consume s and he design o in e es a e ules. J Money, C edi ,
Banking 36(4):739-764
Galí J, López-Salido JD, Vallés J (2007) Unde s anding he e ec s o go e nmen spending on consump ion. JEEA 5(1):227-
270
Ga cía A, Ramajo J (2005) Fiscal policy and p i a e consump ion beha iou : he Spanish case. Empi ical Econ30:115-135
Gio dano R, Momigliano S, Ne i S., Pe o i R (2007) The e ec s o iscal policy in I aly: e idence om a VAR model. Eu op
J Poli Economy 23(3):707-733
Hall P (1992) The boo s ap and Edgewo h expansion. Sp inge , New Yo k
Hamil on JD (1994) Time se ies analysis. Uni e si y P ess P ince on, New Je sey
Hauzenbe ge K (2010) Flexible egime-swi ching p ojec ions o es ima e he dynamic e ec s o a go e nmen spending
s imulus. Eu opean Uni e si y Ins i u e wp, Flo ence
Johansen, S. (1995). Likelihood-based in e ence in coin eg a ed ec o au o eg essi e models. Ox o d Uni e si y P ess.
King RG, Plosse CI, S ock JH, Wa son M (1991) S ochas ic ends and economic luc ua ions. Ame Econ Re 81(4):819-
840
K usec D (2003) The e ec s o iscal policy on ou pu in a s uc u al VEC model amewo k: he case o ou EMU and ou
non-EMU OECD coun ies. Eu opean Uni e si y Ins i u e wp, Flo ence
Lü kepohl H (1991) In oduc ion o mul iple ime se ies analysis. Be lin: Sp inge -Ve lag
Lü kepohl H, K ä zig M (2004) Applied ime se ies econome ics. Camb idge Uni P
Mankiw NG (2000) The sa e s-spende s heo y o iscal policy. Ame Econ Re 90:120-125
Mi nik S., Neuman T (2001) Dynamic e ec s o public in es men : ec o au o eg essi e e idence om six indus ialized
coun ies. Empi ical Econ 26:429-446
16
Monacelli T, Pe o i R (2010) Fiscal policy, he eal exchange a e and aded goods. Economic J 120(544):437-461
Pe o i R (2004) Es ima ing he e ec s o iscal policy on OECD coun ies. IGIER-Bocconi wp
Pe o i R (2007) In sea ch o he ansmission mechanism o iscal policy. NBER wp 13143
Pesa an MH, Shin Y (2002) Long- un s uc u al modelling. Econome Re 21(1):49-87
Ramos X, Roca-Sagales O (2008) Long- e m e ec s o iscal policy on he size and dis ibu ion o he pie in he UK. Fisc
s ud 29(3):387-411
Ra n M, Schmi -G ohè S, U ibe M (2004) Deep habi s. Re Econ S ud 73:195-218
Sawye M (2007) Fiscal policy unde New Labou . Camb idge J Econ 31:885-899
Wang B (2005) E ec s o go e nmen expendi u e on p i a e in es men : Canadian empi ical e idence. Empi ical Econ
30:493-504
Wilcox D (1989) The sus ainabili y o go e nmen de ici s: implica ions o he p esen alue bo owing cons ain . J Money,
C edi , Banking 21:291-306
Figu es
Figu e 1: The h ee componen s o public spending in he UK
Sou ce: OECD qua e ly Economic Ou look 83, own compu a ions.
18
Figu e 2: The da a
Sou ce: OECD qua e ly Economic Ou look no. 83, au ho s' calcula ions.
Figu e 3: SVEC impulse esponses – o al go e nmen spending model
No es: 95% con idence bands a e boo s apped using he Hall me hod wi h 500 eplica ions.
Figu e 4: SVEC impulse esponses – go e nmen wages model
No es: 95% con idence bands a e boo s apped using he Hall me hod wi h 500 eplica ions.

20
Figu e 5: SVEC impulse esponses – ne social secu i y expendi u e model
No es: 95% con idence bands a e boo s apped using he Hall me hod wi h 500 eplica ions.
Figu e 6: SVEC impulse esponses – go e nmen consump ion model
No es: 95% con idence bands a e boo s apped using he Hall me hod wi h 500 eplica ions.
Tables
Table 1: Uni oo es s o sample pe iod 1981:1 o 2007:4
Va iable
De e minis ic e ms
Lags
Tes s a is ic
5% c i . alue
GW
C, T
n(AIC)=n(HQ)=5
-1.3688
-3.41
ΔGW
C
n(AIC)=n(HQ)=4
-3.862**
-2.86
GSS
C
n(AIC)=5
-2.43
-2.86
C
n(HQ)=0
-1.62
-2.86
ΔGSS
C
n(AIC)=n(HQ)=0
-10.72**
-2.86
GC
C, T, SD
n(AIC)=2
-1.71
-3.41
C, T, SD
n(HQ)=1
-1.94
-3.41
ΔGC
C, T
n(AIC)=n(HQ)=1
-10.32**
-3.41
GTOT
C, T
n(AIC)=8
-3.48*
-3.41
n(HQ)=5
-2.9
-3.41
ΔGTOT
C
n(AIC)=4
-3.01*
-2.86
n(HQ)=0
-12.95**
-2.86
T
C, T
n(AIC)=n(HQ)=1
-1.88
-3.41
ΔT
C
n(AIC)=n(HQ)=0
-15.47**
-2.86
C
C, T
n(AIC)=n(HQ)=3
-2.59
-3.41
ΔC
C
n(AIC)=2
-3.71**
-2.86
C
n(HQ)=1
4.50**
-2.86
B
C
n(AIC)=n(HQ)=1
-1.56
-2.86
C, T
n(AIC)=n(HQ)=1
-2.27
-3.41
ΔB
C
n(AIC)=n(HQ)=0
-3.56**
-2.86
No es: C - cons an , T - linea ime end, SD - seasonal dummies. n(AIC) and n(HQ) a e he lag o de s ecommended by he
AIC and HQ c i e ia, espec i ely, when he maximum lag o de is se o 8. ADF c i ical alues om Da idson and
MacKinnon (1993). * and ** indica e ha he null hypo hesis o a uni oo can be ejec ed a he 5% and 1%, espec i ely.
22
Table 2: Coin eg a ion es s o sample pe iod 1981:1 o 2007:4
Va iables
De e minis ic
e ms
H0 ank
Johansen ace es
5% c i . alue
n(AIC)
n(HQ)
GTOT , T , C , B
C, T, SD
0
90.95***
90.95***
63.66
1
45.03**
45.03**
42.77
2
20.26
20.26
25.73
3
7.46
7.46
12.45
[n=2]
[n=2]
GW , T , C , B
C, T, SD
0
83.03***
82.21***
63.66
1
39.31
39.66
42.77
2
12.19
12.81
25.73
3
4.09
3.91
12.45
[n=3]
[n=2]
GSS , T , C , B
C, T, SD
0
94.65***
98.58***
63.66
1
51.75***
50.26***
42.77
2
20.38
18.87
25.73
3
3.96
5.1
12.45
[n=3]
[n=2]
GC , T , C , B
C, T, SD
0
84.05***
84.05***
63.66
1
41.21*
41.21*
42.77
2
17.87
17.87
25.73
3
7.26
7.26
12.45
[n=2]
[n=2]
No es: c - cons an , - linea ime end, sd - seasonal dummies. n(AIC) and n(HQ) indica e ha he lag o de s a e chosen by
he AIC and HQ c i e ion, espec i ely, wi h maximum lag o de se o 8. The ac ual o de s a e gi en in b acke s unde nea h
he es alues. *, ** and *** indica e ha he null hypo hesis can be ejec ed a he 10%, 5% and 1% le els, espec i ely,
based on he c i ical alues om Johansen (1995).
Table 3. Es ima ed coin eg a ing ma ix β and loading coe icien s
Speci ica ion
Loading coe icien s
Fiscal Rules
α1C
α1T
α1G
α1B
α2C
α2T
α2G
α2B
G = GTOT
-
-0.33*
-
-0.08**
T = 0.19G + 0.24***B
0.00***
-0.22**
0.00***
-0.64***
C = -0.12G - 0.12***B
G = GW
0.02*
-0.36***
0.04*
-0.08
T = 0.67***G + 0.26***B
0.11***
-0.24**
-
-0.60***
C = -0.30***G - 0.14***B
G = GSS
-
-0.49***
0.04*
0.18***
T = -0.41***G + 0.31***B
0.00***
-0.56***
-0.03
-0.18***
C = 0.20***G - 0.17***B
G = GC
-
-0.35***
-0.03
-
T = 0.09G + 0.25***B
0.07**
-0.26***
0.01
-0.53***
C = -0.07G - 0.12***B
No es: *, ** and *** indica e ha he null hypo hesis can be ejec ed a he 10%, 5% and 1% le els, espec i ely.
