Is Mone a y Policy in New Membe s
S a es Asymme ic?
Bo ek Vasicek
10.10
De
p
a amen d'Economia A
p
licada
Facul a d'Economia i Emp esa
Aques documen pe any al Depa amen d'Economia Aplicada.
Da a de publicació :
Depa amen d'Economia Aplicada
Edi ici B
Campus de Bella e a
08193 Bella e a
Telè on: (93) 581 1680
Fax:(93) 581 2292
E-mail: [email p o ec ed]
h p://www.ecap.uab.es
Desemb e 2010
Is Mone a y Policy in New Membe s S a es Asymme ic?
Bořek Vašíček
∗
∗∗
∗
Depa men o Applied Economics
Uni e si a Au ònoma de Ba celona
Abs ac
Es ima ed Taylo ules became popula as a desc ip ion o mone a y policy conduc . The e a e
nume ous easons why eal mone a y policy can be asymme ic and es ima ed Taylo ule
nonlinea . This pape es s whe he mone a y policy can be desc ibed as asymme ic in h ee
new Eu opean Union (EU) membe s ( he Czech Republic, Hunga y and Poland), which apply
an in la ion a ge ing egime. Two di e en empi ical amewo ks a e used: (i) a Gene alized
Me hod o Momen s (GMM) es ima ion o models ha allow disc imina ion be ween he
sou ces o po en ial policy asymme y bu a e condi ioned by speci ic unde lying ela ions
(Dolado e al., 2004, 2005; Su ico, 2007a,b); and (ii) a lexible amewo k o sample spli ing
whe e nonlinea i y en e s ia a h eshold a iable and mone a y policy is allowed o swi ch
be ween egimes (Hansen, 2000; Cane and Hansen, 2004). We ind gene ally li le e idence
o asymme ic policy d i en by nonlinea i ies in economic sys ems, some e idence o
asymme ic p e e ences and some in e es ing e idence on policy swi ches d i en by he
in ensi y o inancial dis ess in he economy.
Keywo ds: mone a y policy, in la ion a ge ing, nonlinea Taylo ules, h eshold es ima ion
JEL Classi ica ion: C32, E52, E58
∗ Co espondence Add ess: Depa men o Applied Economics, Uni e si a Au ònoma de Ba celona, Edi ici B,
08193 Bella e a (Ce danyola del Vallès), Spain. E-mail: bo ek. [email protected], bo e[email p o ec ed].
Tel.: +34-93 5814579/1680. Fax: +34-935812292.
1
1. In oduc ion
Since Taylo ’s (1993) in luen ial pape he e has been as esea ch abou he way cen al
banks handle in e es a e se ing. Cla ida e al. (1998, 2000) p opose ha cen al banke s a e
p oac i e a he han eac i e and se in e es a es wi h espec o expec ed alues o
mac oeconomic a iables. Es ima ed mone a y policy ules ypically ake a linea o m
assuming ha mone a y policy esponds symme ically o economic de elopmen s. A
heo e ical unde pinning o a linea policy ule is he linea -quad a ic (LQ) ep esen a ion o
mac oeconomic models wi h an economic s uc u e assumed o be linea and he policy
objec i es o be symme ic ( he loss unc ion is quad a ic).
Howe e , when he assump ions o he LQ amewo k a e elaxed, he op imal mone a y
policy can be asymme ic, which can be ep esen ed by a nonlinea mone a y policy ule.1
The i s sou ce o policy asymme y lies in he nonlinea i ies in he economic sys em. A
common example o such nonlinea i y is a s eepe in la ion-ou pu ade-o when he ou pu
gap is posi i e. The con exi y o he Phillips cu e (PC) implies ha he in la iona y e ec s
o excess demand a e la ge han he disin la iona y e ec s o excess supply (e.g. Lax on e
al., 1999). This can lead op imizing cen al banke s o beha e asymme ically (Dolado e al.,
2005). Howe e , asymme ic mone a y policy can also be ela ed o genuinely asymme ic
p e e ences o cen al banke s. While cen al banks in he pas we e p one o in la ion bias
due o a p e e ence o high employmen o unce ain y abou i s na u al le el (Cukie man,
2000), epu a ion easons can d i e cen al banks, especially hose pu suing in la ion-
a ge ing, o ha e an an i-in la ion bias, which means ha hey espond mo e ac i ely when
in la ion is high o exceeds i s a ge alue (Ruge-Mu cia, 2004). In any case, i seems
plausible ha eal mone a y policy conduc is oo complex o be desc ibed by a simple linea
equa ion and nonlinea ep esen a ion o he mone a y policy can be mo e app op ia e
i espec i e o i s unde lying sou ces.
Nume ous empi ical s udies ha e p o ided e idence ha mone a y policy se ing o many
cen al banks can eally be cha ac e ized as asymme ic. The asymme ic loss unc ion was
ound o a ec he decisions o he Bank o England (Taylo and Da adakis, 2006) and he
1 Asymme ic mone a y policy implies ha mone a y policy ule o he Taylo ule, which is a schema iza ion o
policy eac ion unc ion, is nonlinea .
2
US Fed (Dolado, e al., 2004). Bec e al. (2002) con i m ha he US Fed, Bundesbank and he
Bank o F ance esponded mo e ac i ely o in la ion du ing economic booms. Leu and Sheen
(2006) and Ka adelinkli and Lees (2007) de ec an asymme ic esponse o he ou pu gap by
he Rese e Bank o Aus alia. Su ico (2007a) claims ha he Eu opean Cen al Bank ECB
esponded in i s ea ly yea s mo e s ongly o ou pu con ac ion han expansions and ha he
le el o he in e es a e i sel was a sou ce o policy asymme y. Su ico (2007b) es ablishes
simila e idence on he FED’s asymme ic esponse o he ou pu gap in he p e-Vocke e a
and quan i ies he in la ion bias induced by such a policy. The asymme ies due o he
con exi y o he PC ound in some Eu opean coun ies (Dolado e al., 2005) and he ECB
(Su ico, 2007a) we e linked o wage igidi y in he Eu opean coun ies.
The mone a y policy in he new membe s a es (NMS) o he EU was subjec o subs an ial
changes along hei economic ansi ion. They we e expe imen ing wi h di e se mone a y
policy and exchange a e egimes un il he la e 1990s when he policy egimes we e se led in
line wi h he bipola iew. Some coun ies adop ed ha d exchange a e pegs, which pu a
signi ican cons ain on hei mone a y policy ( he Bal ic S a es, Bulga ia, Cyp us, Mal a),
while o he economies decided o main ain an o e all lexible exchange a e, allowing hei
cen al banks o pu sue in e nal mac oeconomic a ge s, in pa icula p ice s abili y (Cen al
Eu opean Coun ies, Romania).
The empi ical e idence on he mone a y policy se ing in he NMS is a he limi ed. A ew
s udies (Ma ía-Dolo es, 2005; F ömmel and Schobe , 2006; Mohan y and Klau, 2007;
Vašíček, 2010a) p o ided some e idence on linea mone a y policy ules. Howe e , some
na a i es sugges ha mone a y policy in he NMS can be asymme ic. In pa icula ,
coun ies who adop ed a egime o di ec in la ion a ge ing (DIT) can show asymme ic
beha io due o easons o epu a ion. Ho á h (2008) inds some e idence o an asymme ic
policy o he Czech Na ional Bank a e i adop ed his egime. The eason was he need o
gain c edibili y and o ancho in la ion expec a ions. On he o he hand, DIT is lexible
enough o allow he policy make s no o con ac demand when in la ion is sligh ly abo e he
a ge and he shocks a e likely o be sho -li ed (Blinde , 1997). Simila ly, i seems plausible
ha o he conce ns such as economic g ow h and inancial s abili y can lead o he empo al
dismissal o in la ion a ge s.

3
In his pape , we es he hypo hesis o asymme ic mone a y policy in h ee Cen al Eu opean
NMS, he Czech Republic, Hunga y and Poland, who adop ed he amewo k o DIT and
main ain a lexible exchange a e. We employ wo empi ical amewo ks o es he policy
asymme y: (i) a amewo k based on an unde lying s uc u al model, which allows
disc imina ion be ween he sou ces o policy asymme y bu is condi ioned by he speci ic
model se ing; and (ii) a lexible econome ic amewo k, whe e mone a y policy swi ches
be ween wo egimes, acco ding o a h eshold a iable. Besides he common choice o he
h eshold a iable, in la ion de ia ion om he a ge and he s ance o he business cycle, we
use also a deg ee o inancial s ess in he economy o see whe he in la ion- a ge ing cen al
banks beha e di e en ly when he economy is dis essed.
The es o he pape is o ganized as ollows: he nex sec ion b ie ly e iews he main
a ionales o asymme ic mone a y policy; in sec ion 3, we p esen he empi ical s a egies
ha will be used o es he policy asymme y and in sec ion 4, ou da ase . In sec ion 5, we
e iew he empi ical esul s and he las sec ion concludes.
2. Ra ionales o asymme ic mone a y policy
While linea mone a y policy ules can be de i ed in he common LQ amewo k (S ensson,
1999; Cla ida e al. 1999), he nonlinea policy a ises when we allow o some depa u es
om his se ing. The s uc u e o he economy is commonly desc ibed by wo-equa ions,
acking he e olu ion o in la ion and he ou pu :
(
)
[
]
{
}
1 1
1
E g y
π β π βθ π λ ξ
− +
= − + + +
(1)
(
)
[
]
[
]
{
}
1 1 1
1
y y E y i E
µ µ ϕ π ς
−
− + +
= − + − +
(2)
whe e
π
is he in la ion a e,
y
is he ou pu gap,
i
is he nominal sho - e m in e es a e
and
ξ
and
ς
a e supply and demand shock, espec i ely. Eq. (1) ep esen s he AS schedule
o he PC and Eq. (2) is he in e - empo al IS cu e. While he adi ional backwa d-looking
model (S ensson, 1997) assumes
0
β µ
= =
, he New Keynesian model (Cla ida e al., 1999)
is o wa d-looking
1
β µ
= =
. The mone a y au ho i y is usually assumed o se he nominal
in e es a e so as o minimize he loss unc ion:
(
)
{
}
*
, ,
s s
L y x
π π
+ +
= −
(3)
4
whe e
ep esen s gene al unc ional o m, which can be quad a ic i he p e e ences a e
symme ic,
*
π
is he in la ion a ge and
x
a e o he policy objec i es such as exchange a e
s abiliza ion o in e es a e smoo hing.
Fo a de i a ion o asymme ic mone a y policy, which in p ac ice is ep esen ed by a
nonlinea eac ion unc ion, bo h unc ional o ms
and
g
a e impo an .
While Dolado e al.
(2005) assume a case whe e
g
is con ex, Dolado e al. (2004) p opose a mo e gene al se ing
wi h
g
ha may no be linea and
may no be quad a ic, hough bo h pape s use a backwa d-
looking model (
0
β µ
= =
). Su ico (2007a,b) employs a o wa d-looking se ing (
1
β µ
= =
)
wi h a linex o m o he policy loss unc ion, adding addi ional policy objec i es
x
in Eq. (3),
in pa icula ha cen al banks wan s o minimize he in e es a e ola ili y a ound he
implici a ge as well as he de ia ion o he cu en in e es a e om he pas alue.
The e o e, di e en combina ion o unc ional o ms (1)–(3) gi e ise o di e en e sions o
nonlinea policy ule ha can be b ough o he da a. Howe e , imposing a speci ic model
s uc u e can u n p oblema ic gi en ha many a iables and hei ela ions a e no di ec ly
obse able. In addi ion, he NMS a e small open economies, whe e nume ous ex e nal ac o s
may a ec he domes ic in la ion
π
and ou pu
y
and he ela ions i sel can be subjec o
s uc u al change. The e o e, an al e na i e can be o use an empi ical amewo k ha acks
asymme ies in a mone a y policy se ing bu does no ely on he speci ic s uc u e o he
model.
3. Empi ical es ing o asymme ic mone a y policy
The e a e di e se empi ical s a egies o es mone a y policy asymme y. They ypically
consis o an es ima ion o a mone a y policy ule ha includes some nonlinea ea u e. We
de ine, as a benchma k, a linea o wa d-looking mone a y policy ule (Cla ida e al., 1998,
2000), which can also be de i ed as op imal mone a y policy in he New Keynesian model
(Cla ida e al., 1999):
(
)
(
)
(
)
* * *
s s s k
i i E E y
β π π π γ ε
+ + + +
 
= + − Ω − + Ω+
 
 
(4)
whe e all he a iables ha e he p e ious meaning,
*
i
is he in e es a e a ge ,
i
is he
nominal equilib ium in e es a e,
E
is he expec a ion ope a o ,
Ω
is he in o ma ion
5
a ailable o he cen al bank a he ime o policy decision and
ε
is he e o e m. Gi en ha
eal- ime da a, unde lying he policy decision (see O phanides, 2001), is no a ailable o he
NMS, we need o use ac ual ealiza ions o he a iables as p oxies o hei expec ed alues.
In addi ion, we allow o in e es a e smoo hing. The e o e, he obse ed sho - e m in e es
a e is a combina ion o a ule-implied a ge
*
i
and he p e ious alue o he in e es a e
1
i
−
:
( )
(
)
(
)
*
1 12 12
1
i i y
ρ ρ α β π π γ υ
− + +
= + − + − + +
(5)
whe e all he a iables ha e he p e ious meaning,
α
is he cons an e m,
ρ
is he smoo hing
coe icien , ep esen ing he s eng h o policy ine ia, and
υ
is he new e o e m. The
pa ial-adjus men beha io is ypically jus i ied by he ac ha sudden changes in in e es
a e could ha e des abilizing e ec s on inancial ma ke s bu i s ue in ensi y is s ill he
subjec o deba e (Rudebusch, 2002, 2006). We se
s
= 12, which co esponds o common
in la ion a ge ing ho izon, and
k
= 0, assuming ha cen al banks espond o he cu en
ou pu gap. Gi en ha he cu en alue o he po en ial ou pu is no obse able, i mus be
also p oxied by ex-pos da a, which makes i also po en ially an endogenous eg esso . The
e o e m
υ
is a linea combina ion o o ecas e o s o he igh -hand side a iables and he
o iginal exogenous dis u bance
ε
. The e o e, i shall be o hogonal o he p esen in o ma ion
se
Ω
. We will i he Eq. (5) as a benchma k linea model using GMM wi h common
Newey-Wes (1994) co a iance es ima o obus o he e oskeda ici y and au oco ela ion. The
ins umen s a e h ee lags o sho - e m in e es a e, in la ion a e, he ou pu gap and in e es
a e in he eu o a ea.
2
3.1 Nonlinea i ies in he economic sys em
Mone a y policy asymme y can be ela ed o nonlinea i ies in he economic sys em. In
pa icula , nominal p ice s ickiness can cause a nonlinea ade-o be ween in la ion and
ou pu . Dolado e al. (2005) de i e nonlinea mone a y policy ule when he PC is con ex.
They p opose o augmen he s anda d linea policy ule, such as Eq. (5), by an in e ac ion
2 The e is a ce ain con o e sy abou addi ional a iables ha can a ec he in e es a e decisions. In pa icula ,
small open economies could adjus he in e es a e e.g. o he exchange a e o in e na ional in e es a es.
Howe e , he h ee NMS use he DIT, whe e he domes ic p ice s abili y is he only o icial policy a ge .
Mo eo e , he e is no e idence ha Hunga ian and Polish cen al banks espond o any addi ional a iable
(Vašíček, 2010a). Al hough he in e es a e o he eu o a ea u ns some imes signi ican in he es ima ed policy
ule o he Czech Na ional Bank (Ho á h, 2008, Vašíček, 2010a), i is puzzling whe he his means a genuine
aim o s abilize domes ic in e es a e is-à- is he eu o a ea o i is only an e ec o he eu o a ea in e es a e on
he Czech in la ion o ecas , which he cen al bank esponds o. Tha is why we include he eu o a ea in e es
a e as ins umen a he han eg esso .
6
e m o he expec ed in la ion and he ou pu gap gi en ha any in la iona y p essu es d i en
by he ou pu gap a e la ge i he PC is con ex, which calls o an addi ional in e es a e
inc ease whene e he ou pu gap is posi i e.
To implemen empi ically his amewo k, we es ima e in he i s s ep a e y simple
backwa d-looking PC de ined as:
2
1 1 1
y y u
π α βπ γ γφ
− − −
= + + + +
(6)
whe e he p esen in la ion a e
π
depends on i s lagged alue
1
π
−
and a lagged ou pu gap
1
−
y
. The PC is nonlinea when he coe icien
φ
is signi ican ly di e en om ze o, in
pa icula i is con ex when
φ
>0 and conca e when
φ
<0. Second, we es ima e he
co esponding nonlinea policy ule:
( )
(
)
(
)
(
)
* *
1 12 12 12 12
1
i i y y
ρ ρ α β π π γ κ π π υ
− + + + +
= + − + − + + − +
(7)
whe e he posi i e and s a is ically signi ican alue o he coe icien accompanying he
in e ac ion e m o in la ion and ou pu
κ
is an e idence o ule asymme y. In pa icula , he
inc ease o he in e es a e is mo e han p opo ional when he in la ion is abo e he de ined
a ge o he ou pu gap is posi i e.
3
3.2 Asymme ic p e e ences o he cen al bank
Asymme ic p e e ences wi h espec o economic ou comes ep esen ano he a ionale why
he cen al banks can beha e asymme ically. They may disp opo ionally dec ease he
in e es a e when he ou pu is below i s po en ial ( o p e en u he ecession) o inc ease i
when he in la ion exceeds he speci ied a ge ( o c edibili y easons). Dolado e al. (2004)
show ha unde asymme ic p e e ence, he op imal policy ule is nonlinea , i espec i e o
he o m o he AS schedule. In hei model when he cen al bank assigns a highe weigh o
posi i e in la ion de ia ions om he a ge , he in la ion ola ili y (condi ional a iance)
becomes an addi ional a gumen in he mone a y policy ule
This claim can be empi ically es ed as ollows. Fi s , i he condi ional in la ion a iance is
ime a ying, he esiduals o he PC (Eq. (6)) shall con ain au o eg essi e condi ional
he e oscedas ici y (ARCH) e ec s. The null hypo hesis o condi ional homoskedas ici y can
be es ed by means o an ARCH LM es . I he null is ejec ed, Eq. (6) can be es ima ed mo e
3 As we allow he in la ion a ge o a y in ime, we use an in e ac ion e m o in la ion gap and he ou pu gap
a he han in la ion a e and he ou pu gap as Dolado e al. (2005).
13
Table 1.: GMM es ima es o he linea mone a y policy ule (Eq. (5))
Coun y α
(cons .)
β
(π +12 - π +12
*
)
γ
(y ) ρ
(i -1) R 2
LB
J-s a .
CZE (in l. a g.) 3.23 1.24 0.53 0.93 0.99 0.00 0.85
(0.52)***
(0.47)*** (0.41) (0.01)***
CZE (in l. a g. end ) 3.31 1.35 0.53 0.93 0.99 0.00 0.85
(0.55)***
(0.51)*** (0.43) (0.01)***
CZE (in l. end) 2.62 1.34 0.35 0.94 0.99 0.00 0.66
(0.55)***
(0.60)** (0.39) (0.01)***
HUN (in l. a g.) 4.37 2.32 3.14 0.97 0.93 0.00 0.68
(6.24)***
(3.16) (3.06) (0.02)***
HUN (in l. a g. end )
-5.79 6.49 6.87 0.98 0.93 0.00 0.45
(22.36) (10.17) (9.20) (0.02)***
HUN (in l. end) 5.77 4.98 6.01 0.98 0.93 0.00 0.68
(3.15)* (4.36) (4.19) (0.01)***
POL (in l. a g.) 5.05 2.43 3.59 0.96 0.99 0.00 0.72
(1.02)***
(1.59) (1.82)* (0.01)***
POL (in l. a g. end ) 5.41 1.35 2.41 0.95 0.99 0.00 0.80
(0.69)***
(1.20) (1.22)* (0.01)***
POL (in l. end) -6.84 25.54 21.28 0.99 0.99 0.00 0.38
(11.42)* (20.15) (14.65) (0.01)***
No es: HAC s anda d e o s in pa en hesis. *, **, *** deno es signi icance le els a 10, 5 and 1%. LB is p-
alue o Ljung-Box es o 1. o de se ial co ela ion. J-s a is p- alue o Sa gan o e iden i ica ion es .
5.2 Nonlinea mone a y policy ules due o nonlinea i ies in he economic sys em
The i s po en ial d i e o nonlinea mone a y policy is a con ex AS schedule implying ha
in la iona y endencies a e s onge (due o capaci y cons ains) when he ou pu gap is
posi i e. Hence, as a i s s ep we mus es whe he he e is any e idence on he nonlinea
ela ion be ween he in la ion a e and he ou pu gap.
Es ima es o he linea and nonlinea e sion o he simple backwa d-looking PC (Eq. (6))
appea in Table 2. Besides OLS we also use a GARCH(1,1) model o ake in o accoun he
po en ial ime- a ying ola ili y o in la ion. We a e mainly in e es ed in sign and s a is ical
signi icance o he coe icien o he squa ed ou pu gap
γφ
. The PC is con ex when his e m
is posi i e. The esul s show ha he e is li le e idence o any (linea o nonlinea )
ela ionship be ween in la ion and he s ance o business cycle in hese h ee NMS. This is
also e iden by simple isual inspec ion o Figu e 1, showing he sca e plo s be ween he
smoo hed in la ion a e
(
)
1
ˆ
π βπ
−
− and he ou pu gap
1
y
−
. Al hough he esul s can be
a ec ed by noise in measu ing he ou pu gap, he e is also some e idence showing ha

14
in la ion a es in he NMS ha e signi ican ex e nal de e minan s (S a e , 2009; Vaší
č
ek,
2010b).
Table 2.: OLS/GARCH es ima es o simple linea /non-linea Phillips cu es (Eq. (6))
Coun y α
(cons .) β
(π -1) γ
(y -1)
γφ
(y -1
2) ω
(cons .)
ν1
(ξ -1
2
)
ν2
(σ -1
2) R 2
LB
CZE (OLS.) 0.10 0.96 0.06 0.95 0.00
(0.07) (0.01)***
(0.04)
CZE (GARCH)
0.16 0.94 0.02 0.05 -0.04 0.87 0.95 0.00
(0.11) (0.04)***
(0.06) (0.01)***
(0.02)** (0.03)***
CZE (OLS) 0.00 0.97 -0.07 0.04 0.95 0.00
(0.11) (0.02)***
(0.06) (0.04)
CZE (GARCH)
0.06 0.93 -0.02 0.07 0.32 0.95 0.05 0.95 0.00
(0.13) (0.04) ***
(0.07) (0.04) (0.11)***
(0.01)***
(0.03)
HUN (OLS) 0.26 0.95 0.05 0.98 0.00
(0.11)***
(0.01)***
(0.03)
HUN (GARCH)
0.30 0.94 0.05 0.05 -0.02 0.83 0.97 0.00
(0.12)***
(0.02)***
(0.03) (0.06) (0.02) (0.23)***
HUN (OLS) 0.32 0.95 0.05 -0.03 0.97 0.00
(0.11)***
(0.01)***
(0.03) (0.01)***
HUN (GARCH)
0.36 0.95 0.05 -0.03 0.06 -0.03 0.80 0.98 0.00
(0.12)***
(0.02) ***
(0.03) (0.02) (0.06) (0.03) (0.22)***
POL (OLS) 0.10 0.96 0.06 0.98 0.00
(0.07) (0.01)***
(0.04)
POL (GARCH)
0.19 0.93 0.09 0.00 -0.03 1.01 0.98 0.00
(0.04)***
(0.00)***
(0.02)*** (0.00) (0.00)***
(0.00)***
POL (OLS) 0.07 0.96 0.06 0.03 0.98 0.00
(0.08) (0.01)***
(0.04) (0.03)
POL (GARCH)
0.14 0.93 0.09 0.04 0.06 -0.03 0.80 0.98 0.00
(0.07)* (0.02)***
(0.04)*** (0.03) (0.06) (0.03) (0.22)***
No es: S anda d e o s in pa en hesis. *, **, *** deno es signi icance le els a 10, 5 and 1%. LB is p- alue o Ljung-
Box
es o . o de se ial co ela ion.
Figu e 1: Sca e plo s be ween smoo hed in la ion a e and ou pu gap ( he Phillips cu e)
and i ed linea end
-4
-3
-2
-1
0
1
2
3
4
-4 -3 -2 -1 0 1 2 3
GDPCZEGAP(-1)
CPICZE-0.96*CPICZE(-1)
-2
-1
0
1
2
3
-8 -6 -4 -2 0 2 4 6
GDPPOLGAP(-1)
CPIPOL-0.96*CPIPOL(-1)
-2
-1
0
1
2
3
4
5
-6 -4 -2 0 2 4 6
GDPHUNGAP(-1)
CPIHUN-0.95*CPIHUN(-1)
15
Al hough he p e ious esul s pu in o ques ion he con exi y o he AS schedule, we con inue
o es ima e Eq. (7), whe e he in la ion-ou pu in e ac ion e m appea s as an addi ional
eg esso . These esul s a e epo ed in Table 3. As expec ed, his e m is mos ly insigni ican
and he e is no indica ion o asymme ic cen al bank eac ion d i en by a nonlinea PC. In
any case, i is impo an o keep in mind ha he esul s a e condi ioned by he unde lying
model.
12
Table 3.: GMM es ima es o he nonlinea mone a y policy ule (Eq. (7))
Coun y α
(cons .)
β
(π +12 - π +12
*
)
γ
(,y ) ρ
(i -1)
κ
( (π +12 - π +12
*)y )
R 2
LB
J-s a .
CZE (in l. a g.) 3.40 1.57 0.56 0.93 -0.48 0.99 0.00 0.72
(0.49)***
(0.47)*** (0.32)* (0.02)*** (0.38)
CZE (in l. a g. end ) 3.70 1.80 0.57 0.95 -0.68 0.99 0.00 0.59
(0.48)***
(0.45)*** (0.33)* (0.01)*** (0.36)*
CZE (in l. end) 1.73 2.39 1.47 0.94 -1.06 0.99 0.00 0.43
(0.83)** (0.86)*** (0.59)** (0.01)*** (0.59)*
HUN (in l. a g.) -0.86 3.64 8.41 0.98 -3.55 0.93 0.00 0.82
(19.00) (7.90) (12.87) (0.03)*** (5.92)
HUN (in l. a g. end )
1.82 2.64 5.61 0.96 -3.55 0.91 0.07 0.83
(10.74) (4.43) (6.31) (0.03)*** (5.92)
HUN (in l. end) 6.94 2.07 3.69 0.95 -1.71 0.93 0.02 0.71
(1.76)***
(1.95) (1.89)* (0.02)*** (1.40)
POL (in l. a g.) 5.03 2.46 3.61 0.96 0.03 0.99 0.00 0.63
(1.36)***
(1.86) (1.88)* (0.02)*** (1.68)
POL (in l. a g. end ) 7.24 -1.33 2.48 1.03 -3.65 0.98 0.00 0.26
(1.44)***
(1.24) (1.37)* (0.03)*** (2.29)
POL (in l. end) 8.37 -8.08 0.30 1.07 -6.64 0.97 0.00 0.97
(1.43)***
(4.16)** (2.88) (0.05)*** (3.90)
No es: HAC s anda d e o s in pa en hesis. *, **, *** deno es signi icance le els a 10, 5 and 1%. LB is p- alue o Ljung-
Box es
o 1. o de se ial co ela ion. J-s a is p- alue o Sa gan o e iden i ica ion es .
5.3. Nonlinea mone a y policy ules due o asymme ic p e e ences
Cen al banks can espond in a nonlinea way o mac oeconomic a iables due o hei
genuine asymme ic p e e ences. These a e usually ep esen ed by a non-quad a ic loss
unc ion.
Fi s , we explo e whe he he cen al banks o he h ee NMS applied nonlinea policy ule
due o highe weigh assigned o posi i e de ia ion o expec ed in la ion om he a ge .
Dolado e al. (2004) sugges ed acking such nonlinea i y by he inclusion o condi ional
in la ion a iance (Eq. (9)) o an o he wise linea policy ule. The e o e, i s , we need o
12 Mo eo e , his amewo k implici ly assumes ha he h eshold alue o in la ion and ou pu gaps d i ing
policy asymme y a e each ze o because he in e ac ion e m u ns posi i e when in la ion gap and he ou pu gap
a e bo h posi i e o nega i e.
16
check whe he he in la ion ola ili y is uly ime- a ying o be used as an eg esso in Eq. (9)
. The in la ion is again modeled by simple backwa d-looking PC (Eq. (6)) and he ARCH LM
es is used o check he neglec ed ARCH in esiduals. The es gi es a i ma i e e idence o
he Czech Republic and Poland bu canno ejec he null o no condi ional he e oskedas ici y
o Hunga y. Condi ioned on hese esul s, we e-es ima e he PC using GARCH (1,1). The
esul s o he co esponding mean and a iance equa ion bo h o linea and quad a ic
specia ion o he PC appea in Table 2. We can see ha he condi ional a iance o in la ion is
a a he pe sis en p ocess in he h ee coun ies as he coe icien o he GARCH e m
ν2
is
signi ican and close o uni y. We ob ain he es ima ed se ies o condi ional in la ion a iance
and use i as a eg esso (Eq. (9). The esul s appea in Table 4.
Table 4.: GMM es ima es o he nonlinea mone a y policy ule (Eq. (9))
Coun y α
(cons .)
β
(π +12 - π +12
*
)
γ
(,y ) ρ
(i -1)
κ
(σ
2) R 2
LB
J-s a .
CZE (in l. a g.) 2.63 1.35 1.19 0.92 4.72 0.99 0.00 0.59
(0.80)***
(0.38)*** (0.34)*** (0.01)*** (1.97)**
CZE (in l. a g. end ) 2.74 1.70 1.58 0.91 5.93 0.99 0.00 0.93
(0.77)***
(0.48)*** (0.41)* (0.01)*** (1.57)***
CZE (in l. end) 0.13 -0.53 3.85 0.99 11.31 0.99 0.00 0.77
(6.89) (2.62) (6.55)** (0.02)*** (18.82)
HUN (in l. a g.) 27.54 0.75 0.72 0.94 -71.01 0.93 0.00 0.54
(8.39)***
(1.06) (0.85) (0.02)*** (33.66)**
HUN (in l. a g. end )
40.07 1.88 1.56 0.95 -125.81 0.93 0.00 0.37
(14.15) (1.86) (0.29) (0.02)*** (60.94)**
HUN (in l. end) 6.94 2.07 3.69 0.94 -1.71 0.93 0.02 0.33
(1.76)***
(1.95) (1.89)* (0.01)*** (1.40)
POL (in l. a g.) 4.10 1.53 1.76 0.95 8.44 0.99 0.00 0.80
(1.54)***
(0.47) (0.84)** (0.02)*** (15.41)
POL (in l. a g. end ) 5.05 1.17 1.72 0.95 2.30 0.99 0.00 0.84
(1.67)***
(1.35) (0.80)** (0.02)*** (16.53)
POL (in l. end) 4.26 1.08 2.68 0.94 9.16 0.99 0.01 0.68
(1.26)***
(1.57) (0.85)*** (0.02)*** (11.10)
No es: HAC s anda d e o s in pa en hesis. *, **, *** deno es signi icance le els a 10, 5 and 1%. LB is p- alue o Ljung-
Box es o 1. o de se ial co ela ion. J-s a is p- alue o Sa gan o e iden i ica ion es .
The sho - e m in e es a e esponds signi ican ly o he condi ional in la ion a iance in he
Czech Republic, which sugges s ha he Czech Na ional Bank handles he in la ion in an
asymme ic manne , in pa icula ha i weigh s mo e posi i e de ia ions om he a ge han
nega i e ones. On he con a y, he condi ional in la ion a iance en e s wi h a coun e -
in ui i e nega i e sign o Hunga y, which is likely ela ed o noisiness ( he esiduals o PC
o Hunga y does no con ain ARCH e ec s) and e y low a iance o his se ies (s anda d
de ia ion is 0.02 as compa ed 0.43 o he Czech Republic and 0.16 in Poland). In any case,
17
he esul s a e again condi ioned by he speci ica ion o PC ha was used o de i e he
condi ional in la ion a iance.
An al e na i e way o es whe he mone a y policy ule is nonlinea due o asymme ic
p e e ences is sugges ed by Su ico (2007a,b). His app oach does no equi e es ima ion o
condi ional in la ion a iance o es asymme ic esponse o in la ion. In addi ion, i allows
he es ing o whe he he cen al bank has asymme ic p e e ences wi h espec o he ou pu
gap and he in e es a e gap, he la e is de ined as a de ia ion o he cu en in e es a e
om i s long- e m equilib ium alue. The asymme ic p e e ences en e he policy ule by
squa e componen s o in la ion, ou pu and in e es a e gaps (Eq. (10)). We adjus he
nonlinea ule de i ed in Su ico (2007a,b) o be mo e plausible o he in la ion- a ge ing
NMS. In pa icula , we eplace he esponse o con empo aneous in la ion by esponse o
expec ed in la ion gap gi en ha in la ion- a ge ing cen al banks a e o wa d-looking and he
in la ion a ge is no cons an . The es ima es o such nonlinea policy ule appea in Table 5.
The columns wi h es ima es o
κ1
,
κ2
and
κ4
e e o nonlinea i ies ela ed o asymme ic
p e e ences o in la ion, ou pu and in e es a e gaps, espec i ely, and
κ3
cap u es a esponse
o nonlinea i ies in economic s uc u e. Fi s , he only coun y whe e we ind some e idence
o asymme ic esponse o he in la ion gap is Hunga y, hough he sign o coe icien
κ1
is
nega i e, implying a s onge esponse when in la ion is below i s a ge . This con a-in ui i e
inding is in ac consis en wi h he e idence om Eq. (9) whe e we ind a nega i e esponse
o condi ional in la ion a iance. On he o he hand, we do no con i m he p e ious inding
ha he Czech Na ional Bank ea ed he posi i e in la ion de ia ion asymme ically om i s
a ge . Second, he coe icien
κ2
o he squa ed ou pu gap is insigni ican o he h ee
coun ies and i hei cen al banks conside ed he s ance o business cycle (see Tables 2–4),
hey did i in a symme ic manne . Thi d, o all coun ies we e eal a p e e ence o limi he
ola ili y o he cu en in e es a e om i s equilib ium alue (p oxied by he in e cep
α
).
The posi i e alue o
κ4
, ound in he Czech Republic and Hunga y, e lec s a dis as e o
ac ual in e es a es exceeding he equilib ium alue. The nega i e alue ound o Poland can
be a sign ha he Polish Na ional Bank was esis an o keeping he in e es a es oo low. In
ac , he p e e ence o highe in e es a es (
κ4
nega i e) can also be an indica ion o a
p e e ence o p ice s abili y, while he opposi e (
κ4
posi i e) can also indica e a p e e ence o
a oid con ac ion. As compa ed o he benchma k linea case (Eq. (5)), he in e es a e
smoo hing has subs an ially dec eased o mo e plausible le els (Rudebusch, 2002, 2006) as
compa ed wi h he linea case (Eq. (5)). Finally, he in la ion esponse coe icien
β
is no
18
al e ed o he Czech Republic and Poland bu i u ns signi ican and highe han uni y o
Hunga y, indica ing a s abilizing na u e o mone a y policy conduc when he nonlinea na u e
o mone a y policy is aken in o accoun . These indings a e p omising as compa ed o Su ico
(2007a), who ob ains o he ECB less plausible esul s such as a nega i e and insigni ican
esponse o in la ion a e.
13
Due o easons o space, we do no epo he au oco ela ion and
o e -iden i ica ion es s bu hey p o ide a e y simila pic u e as in p e ious ables.
Table 5.: GMM es ima es o he nonlinea mone a y policy ule (Eq.(10))
Coun y α
(cons .)
β
(π +12 - π +12
*
)
γ
(,y ) ρ
(i -1)
κ 1
(π +12 - π +12
*)2
κ 2
(,y
2)
κ 3
(π +12 - π +12
*)y
κ 4
(i -α )2 R 2
CZE (in l. a g.) 2.26 1.05 0.03 0.85 0.03 0.63 -0.32 0.07 0.99
(1.02)**
(0.47)** (0.52) (0.01)***
(0.18) (0.53) (0.32) (0.02)***
CZE (in l. a g. end )
2.29 1.54 0.21 0.89 -0.12 1.17 -0.39 0.06 0.99
(1.43) (0.72)** (0.91) (0.05)***
(0.37) (0.90) (0.54) (0.03)*
CZE (in l. end) 2.11 0.71 0.12 0.82 0.44 0.00 -0.63 0.04 0.99
(0.88)* (0.46) (0.42) (0.04)***
(0.29) (0.49) (0.31)** (0.02)*
HUN (in l. a g.) 6.59 2.09 -0.13 0.58 -0.55 -0.14 -0.15 0.14 0.88
(0.43)***
(0.71)*** (0.37) (0.16)***
(0.23)** (0.16) (0.33) (0.02)***
HUN (in l. a g. end )
6.31 1.72 0.03 0.59 -0.43 -0.07 0.12 0.13 0.90
(0.36)***
(0.64)*** (0.27) (0.13)***
(0.21)** (0.11) (0.27) (0.01)***
HUN (in l. end) 8.14 2.43 0.81 0.91 -1.49 0.52 -0.79 0.10 0.93
(2.31)***
(2.96) (1.24) (0.06)***
(2.19) (0.91) (1.14) (0.06)
POL (in l. a g.) 19.84 0.58 -0.46 0.72 -0.37 1.38 0.80 -0.08 0.98
(6.59)***
(0.73) (1.18) (0.41)***
(0.35) (2.64) (1.78) (0.04)*
POL (in l. a g. end )
20.26 0.36 -0.49 0.51 -0.35 0.60 0.63 -0.07 0.97
(5.97)***
(0.45) (1.05) (0.63) (0.23) (1.35) (1.36) (0.03)**
POL (in l. end) 4.57 0.28 0.17 0.83 0.78 0.23 2.14 0.06 0.99
(2.16)**
(1.15) (0.75) (0.19)***
(0.74) (1.40) (2.34) (0.03)*
No es: HAC s anda d e o s in pa en hesis. *, **, *** deno es signi icance le els a 10, 5 and 1%. LB is p- alue o Ljung-
Box es o 1.
o de se ial co ela ion. J-s a is p- alue o Sa gan o e iden i ica ion es .
5.4. Nonlinea mone a y policy ules ia h eshold e ec s
As a gued be o e, he p e ious me hods o in e ence on policy asymme y es s on he
speci ic assump ion abou he s uc u e o he economy and he cen al bank’s loss unc ion.
In wha ollows, we use he empi ical o wa d-looking policy ule p oposed by Cla ida e al.
(1998, 2000) and allow he esponse coe icien s o swi ch be ween wo egimes acco ding o
he e olu ion o h eshold a iable. Gi en ha he me hod o h eshold es ima ion (Hansen,
2000; Cane and Hansen, 2004) equi es he h eshold o be exogenous, we use as a h eshold
obse ed ( a he han expec ed) alues.
14
Using in la ion and he ou pu gap as h esholds, we
13 A bi su p isingly, he in e p e s hese esul s as e idence ha he ECB ollows a nonlinea policy ule.
14 The econome ic p ocedu e is no sui able i he a iables ha e a uni oo . We apply common es s o uni
oo s, ejec ing i (a con en ional signi icance le els) o all he ime se ies used o es ima ion.

19
wan o see whe he he in e es a e se ing di e s in high and low in la ion egimes and in
ecession and expansions. Mo eo e , we include a new a iable ha can a guably gi e some
insigh on asymme ies in a mone a y policy se ing, he inancial s ess index (EM-FSI). In
his case, we y o unco e whe he cen al banks al e hei conside a ion o common policy
a ge s in he ace o inancial ins abili y and whe he hey di ec ly adjus policy a es
acco ding o he deg ee o inancial s ess in he economy.
The in e ence on mone a y policy asymme y has so a been ca ied ou by means o
con en ional - es s o s a is ical signi icance o addi ional nonlinea e ms (in la ion – ou pu
in e ac ion e m, condi ional in la ion a iance o squa ed e ms o in la ion, ou pu and
in e es a e gaps). Wi h he cu en me hod, he policy asymme y is es ed by means o
h eshold e ec s. Un o una ely, a s anda d Wald es compa ing he poin es ima es in each
egime canno be used because he me hod p o ides a sample spli e en in he absence o ue
h eshold e ec s, which makes es ima es inconsis en .
15
Gi en ha he h eshold es ima ion is
based on he minimiza ion o he squa ed esidual o Eq. (14), we can d aw he in e ed LR
s a is ics (Eq.(16)) o he en i e se o possible h eshold alues
»
o e alua e he p ecision
o he es ima ed h eshold (see Figu es A.3 – A.5).
(
)
n
LR Q
eaches i s minimum, ze o, a he
es ima ed h eshold
ˆ
Q
. The ho izon al line ep esen s he con idence in e al and he alues
o Q whose
(
)
n
LR
Q
a e below his line a e wi hin he con idence in e al. The shape o
(
)
n
LR
Q
indica es he s eng h o he h eshold e ec . I he sequence o
(
)
n
LR
Q
is peaked
wi h a clea ly de ined minimum (o o m V), i is also an indica ion o a signi ican h eshold
e ec , which jus i ies sample spli ing and sepa a e es ima ion o each subsample. On he
con a y, i egula shape whe e
(
)
n
LR
Q
c osses he con idence in e al mo e han once and
he minimum is less e iden , is an indica ion ha he sample may be spli mo e han once o
ha he e is no h eshold e ec a all.
15 The me hod spli s he sample a he alue o h eshold a iable ha minimizes he esiduals o Eq. (14). When
he spli s imply ha one egime con ains only a minimum possible numbe o obse a ions (10% o he o al
sample), while he o he he emaining majo i y, i is an indica ion o no well-de ined h eshold. The Wald es
compa ing he slope es ima es in each egime canno be used as he slope coe icien s in he smalle sub-sample
a e es ima ed e y imp ecisely. In addi ion, wi h no well-de ined h eshold, he es ima ion me hod encoun e s
compu a ion p oblems due o ma ix singula i y. As he h eshold is no iden i ied unde he null hypo hesis o no
h eshold e ec , Hansen (1996) p o ides a boo s apping p ocedu e o es he p esence o he h eshold.
Howe e , gi en unce ain y abou he h eshold a iable, he h eshold alue as well as he numbe o policy
egimes, we assess he p esence o h eshold e ec in ui i ely by he g aphical inspec ion o LR s a is ics
desc ibed below.
20
In Figu es A.3 – A.5, we epo he LR sequence using he in la ion gap, he ou pu gap and
he EM-FSI as al e na i e h eshold a iables. As no ed abo e, we always use he i s lag o
he espec i e a iable as he h eshold a iable mus be exogenous. Fo each h eshold
a iable and coun y, we epo h ee igu es co esponding o a model wi h each measu e o
he in la ion gap. As we can see in Figu e A.3, he h eshold e ec o he in la ion gap is no
e iden and depends on he measu e o he in la ion a ge . Al hough he LR sequences
ea u es a usually well-de ined minimum, i leads o a e y asymme ic sample spli , lea ing
one egime only wi h a minimum numbe o obse a ions pe mi ed (when he in la ion a e
exceeds e y subs an ially he a ge alue o he Czech Republic and Poland and when i is
signi ican ly below i o Hunga y). This disquali ies he easonabili y o he sample spli ing
and asymme ic mone a y policy along he alue o he in la ion gap. The only excep ions
apply o he Czech Republic, when measu ing he in la ion gap by means o in la ion
de ia ion om i s HP end ( igh -mos igu e, es ima ed h eshold is 1.25), and o Poland,
when using in la ion de ia ion om he a ge HP end (middle igu e, es ima ed h eshold is
0.08). Howe e , he es ima ed coe icien s a e mos ly insigni ican in bo h coun ies and
egimes. To sa e space, we do no epo he slope es ima es.
Figu e A.4 plo s he espec i e LR sequences when he ou pu gap is used as he h eshold
a iable. We disca d again he h eshold model o Hunga y as he LR eaches i s minimum
only a e y high alues o he in la ion gap, making he sample spli un easible. Fo he
Czech Republic, we ind a well-de ined h eshold only wi h he in la ion de ia ion om he
end ( he igh -mos panel). In his model, when he ou pu gap exceeds he h eshold alue
(es ima ed a 0.73), i s coe icien
γ
is 2.43 e sus 1.64 when i is below he a ge (in bo h
cases his is highly signi ican ). This sugges s ha he Czech Na ional Bank handles mone a y
policy in an asymme ic way along he business cycle. In pa icula , i is eady o inc ease he
in e es a e by a la ge amoun du ing pe iods o economic expansion.
16
In he i s wo
panels, we can see ha he LR c osses he ho izon al line mo e han once. Howe e , he
sample size does no allow ano he spli . Fo Poland, we ind a p ecise h eshold in he i s
wo models (wi h an in la ion gap de i ed om he ac ual in la ion a ge and om he HP
end o he a ge ). The h eshold alue is es ima ed a -0.05 in bo h cases. While he
co esponding esponse coe icien
γ
is insigni ican in he egime below he h eshold (i.e.
16 These indings mus be in e p e ed wi h cau ion gi en ha he linea mone a y policy ules (Table 1) ea u es a
subs an ially smalle and s a is ically insigni ican esponse o he ou pu gap. Simila ly, he es ima es o
nonlinea policy ules in line wi h Su ico (2007 a,b) epo ed in Table 5, do no indica e s a is ical signi icance o
he squa ed e m o he ou pu gap.
21
when he ou pu is below i s po en ial), i u ns signi ican and eaches a alue o 14 when he
h eshold is b eached. This inding is in e es ing in iew o he linea model es ima es (Table
1.) showing ha he Na ional Bank o Poland (NBP) esponds a he o he ou pu gap han
in la ion. The esul s o he h eshold model sugges ha Polish mone a y policy is
asymme ic along he business cycle. Howe e , his e idence canno be di ec ly in e p e ed
ha he NBP, as a long- e m in la ion a ge e , aims a business cycle s abiliza ion ins ead o
he in la ion a ge . I migh mean ha he ou pu gap a ec s he NBP’s in la ion o ecas ha
is he d i e o in e es a e se ing.
17
A las , we use he inancial s ess indica o (EM-FSI). The e olu ion o his a iable
(no malized o ha e ze o mean) is depic ed in Figu e A.2. I is no able ha all h ee coun ies
expe ienced a deg ee o inancial s ess du ing he ecen global u moil unseen in he
p e ious decade bu ha he s ess was also high as a consequence o he Russian c ises in la e
1998. On he o he hand, unlike many de eloped coun ies, he NMS did no su e an
inc ease in inancial s ess on he e e o he new millennium ollowing he NASDAQ c ash
(2000), he e o is a ack on he US (2001) o he US co po a e scandals (2002). EM-FSI
allows, unlike bina y c isis a iables (Lea en and Valencia, 2008), he in ensi y o inancial
s ess o be measu ed and can be used o he h eshold es ima ion. Ne e heless, i is no
e iden whe he EM-FSI should en e di ec ly in he es ima ed policy ule as eg esso o
“s ay behind” as h eshold a iable d i ing he egime swi ches. In o he wo ds, i is puzzling
whe he he cen al bank may di ec ly espond o some s ess measu e o only o modi y i s
conside a ion o o he objec i es. Consequen ly, we es ima e he h eshold model wi h and
wi hou he inancial s ess as an addi ional eg esso . Figu e A.5 depic s he LR e olu ion
when EM-FSI is included as a eg esso , which is almos iden ical wi h EM-FSI d opped. We
can see ha he h eshold alue is clea ly delimi a ed in all h ee igu es o he Czech
Republic and i s wo igu es o Poland. Fo Hunga y, he LR sequence eaches i s minimum
a e y high alues o he s ess bu he e a e s ill 28 obse a ions in he uppe egime. We
spli all he samples and pu sue GMM es ima ion o each egime. These esul s a e epo ed
in Table 6.
18
17 GARCH es ima es o he Polish PC epo ed in Table 2 indica e ha he ou pu gap has a signi ican e ec on
he in la ion a e.
18 We epo esul s wi h EM-FSI included as an eg esso gi en ha his speci ica ion has a highe i and he
accompanying coe icien o EM-FSI is mos ly s a is ically signi ican .
22
In all bu one case, he uppe egime has subs an ially less obse a ions han he lowe one.
The coe icien o EM-FSI is mos ly signi ican sugges ing ha cen al banke s adjus policy
a es when hey a e aced by inancial s ess. Since he cen al banke s migh espond o
inc easing inancial s ess by mone a y easing, he expec ed sign o he coe icien is nega i e.
Ye , EM-FSI also includes a sub-componen ep esen ing he exchange a e p essu es, in
pa icula domes ic cu ency dep ecia ion,
19
whose p e alence in he o e all index can d i e
an in e es a e inc ease in an a emp o suppo domes ic cu ency. Fo he Czech Republic
and Poland, we ind ha he coe icien
κ
, accompanying EM-FSI is mos ly nega i e and
signi ican when he inancial s ess exceeds he es ima ed h eshold. This sugges s ha bo h
cen al banks dec ease policy a es when he economy su e s high inancial s ess. On he
con a y, he esponse is mos ly insigni ican when he s ess alls below he h eshold alue.
Hunga y seems o be he opposi e case; he in e es a e esponse o inancial s ess is
signi ican ly posi i e and does no di e subs an ially be ween he wo egimes. This could be
ela ed o he o in dep ecia ion p essu es ha we e a signi ican d i e o he Hunga ian
o e all EM-FSI and Hunga ian mone a y policy aced hem by means o in e es a e
inc ease.
20
As a as he o he coe icien s a e conce ned, hei size usually di e s be ween he egimes
wi h he excep ion o he smoo hing pa ame e
ρ
. I s es ima ed size s ill sugges s a subs an ial
deg ee o “policy ine ia” e en when we accoun o possible policy asymme y ia he
h eshold e ec s.
21
On he o he hand, he se ial co ela ion is much less p onounced in he
spli samples han in he models (linea , nonlinea ) based on all obse a ions. The in la ion
coe icien
β
does no ha e any clea pa e n. While wo speci ica ions sugges ha he Czech
Na ional Bank is a s ic e in la ion a ge e when inancial s ess is high, he o he poin s o
he con a y. Fo Poland, in wo speci ica ions, he e is no esponse o in la ion when he
s ess is high and a posi i e esponse when i is low. The hi d speci ica ion ha sugges s he
opposi e pa e n is in ac dubious because i cu s o a ew obse a ions when inancial s ess
is e y low. Fo Hunga y, we s ill canno de e mine he pa e n o i s in la ion a ge ing
19 This subcomponen is no p esen in he inancial s ess index p oposed by he IMF o ad anced economies
(Ca da elli, e al. 2009).
20 Baxa e al. (2010) s udy he esponse o main cen al banks ( he US, he UK, Aus alia, Canada and Sweden)
o inancial s ess using a ime- a ying pa ame e model ha does no impose wo policy egimes bu allows a
unique esponse in each pe iod. Thei esul s also sugges ha he cen al banks a e eady o dec ease policy a es
when he inancial s ess is high. Ne e heless, he size o he esponse a ies subs an ially ac oss coun ies and
ime no excluding pe iods when inancial s ess implied an in e es a e inc ease. Un o una ely, due o he
limi ed leng h o ime se ies a ailable, we canno apply such a amewo k o he NMS.
21 Al hough we ha e ejec ed he p esence o uni oo s in he sho in e es a es, hey a e s ill e y pe sis en a
mon hly equency. This seems o be he main eason o ele a ed policy ine ia ound ac oss his s udy.
29
Figu e A.2: IMF’s Eme ging-Ma ke s Financial S ess Index (EM-FSI)
-6
-4
-2
0
2
4
6
8
10
98 99 00 01 02 03 04 05 06 07 08 09
CZE_EM-FSI
-8
-4
0
4
8
12
16
98 99 00 01 02 03 04 05 06 07 08 09
HUN_EM-FSI
-4
-2
0
2
4
6
8
98 99 00 01 02 03 04 05 06 07 08 09
POL_EM-FSI

30
Figu e A.3: Likelihood a io sequence o di e en alues o he h eshold a iable ( he in la ion gap)
The Czech Republic
Hunga y
Poland
31
Figu e A.4: Likelihood a io sequence o di e en alues o he h eshold a iable ( he ou pu gap)
The Czech Republic
Hunga y
Poland
32
Figu e A.5: Likelihood a io sequence o di e en alues o he h eshold a iable ( he EM-FSI)
The Czech Republic
Hunga y
Poland
TÍTOL
NUM AUTOR DATA
Desemb e
2010
CO2 emissions and economic ac i i y: he e ogenei y
ac oss coun ies and non s a iona y se ies
10.09
Ma ías Piaggio,
Emilio Padilla
Desemb e
2010
Inequali y ac oss coun ies in ene gy in ensi ies: an
analysis o he ole o ene gy ans o ma ion and inal
ene gy consump ion
10.08
Juan An onio Du o,
Emilio Padilla
Desemb e
2010
Is Mone a y Policy in New Membe s S a es
Asymme ic?
10.10
Bo ek Vasicek
Se emb e
2010
How Does Mone a y Policy Change? E idence on
In la ion Ta ge ing Coun ies
10.07
Ja omí Baxa,
Roman Ho á h,
Bo ek Vasícek
Juliol 2010The Wage-P oduc i i y Gap Re isi ed: Is he Labou
Sha e Neu al o Employmen ?
10.06
Ma ika Ka anassou,
Hec o Sala
Juliol 2010Oil p ice shocks and labo ma ke luc ua ions10.05
Ja ie O doñez,
Hec o Sala,
Jose I. Sil a
Juliol 2010Vulne abili y o Po e y: A Mic oeconome ic App oach
and Applica ion o he Republic o Hai i
10.04
E ans Jado e
Maig 2010Nue os y iejos c i e ios de en abilidad que
concue dan con el c i e io del Valo Ac ual Ne o.
10.03
Joan Pasqual,
Emilio Padilla
Ma ç 2010Memo y in Con ac s:
The Expe ience o he EBRD (1991-2003)
10.02
Lionel A ige,
Rosella Nicolini
Feb e 2010Language knowledge and ea nings in Ca alonia10.01
An onio Di Paolo,
Josep Lluís Raymond-Ba a
Desemb e
2009
In la ion dynamics and he New Keynesian Phillips
cu e in EU-4
09.12
Bo ek Vasicek
No emb e
2009
Venezuelan Economic Labo a o y
The Case o he Al uis ic Economy o Felipe Pé ez
Ma í
09.11
Alejand o Aga onow
No emb e
2009
De e minan es del c ecimien o de las emisiones de
gases de e ec o in e nade o en España (1990-2007)
09.10
Vicen Alcán a a Escolano,
Emilio Padilla Rosa
No emb e
2009
He e ogenei y ac oss Immig an s in he Spanish Labou
Ma ke : Ad an age and Disad an age
09.09
Ca ia Nicodemo
No emb e
2009
A sensi i i y analysis o po e y de ini ions09.08
Nicholas T. Long o d,
Ca ia Nicodemo