DLSU Business & Economics Review 28(2) 2019, pp.

30–51

RESEARCH ARTICLE

Measuring Fiscal and Monetary Policies Spillovers

in ASEAN
Thi Mai Lien Dau
School of Development Economics, National Institute of Development Administration, Bangkok, Thailand
daumailien@yahoo.com

Yuthana Sethapramote
National Institute of Development Administration, Bangkok, Thailand

Abstract: This paper aims to comprehensively examine fiscal and monetary policies spillovers to real GDP and inflation
in ASEAN-5 countries. We examine the effects of shocks from each of the ASEAN members and advanced economies by
employing the global vector autoregression (GVAR) model because it allows us to investigate this issue in a multinational
system. The empirical results show several important findings. Generally, both internal and external fiscal and monetary
spillovers have a significant effect on all ASEAN-5 countries, although internal monetary spillover seems to be stronger
than internal fiscal spillovers. At country level, out of the five ASEAN countries, Indonesia’s variables of interest are less
affected by external policy shocks. Regarding fiscal spillovers, expansionary fiscal shocks cause a significant increase in
real GDP of ASEAN-5 countries. Particularity, fiscal spillovers from other East Asia countries (in particular, China) to
ASEAN-5 countries have a much stronger effect than those from the advanced countries. Regarding monetary spillover, the
effects of monetary spillover from advanced economies (especially, the U.S. and European countries) are much larger than
those of the East Asia countries. However, the effects of monetary spillovers on ASEAN countries’ real GDP and inflation
are ambiguous, and we found negative impacts of internal and external monetary spillovers (from some ASEAN countries,
China, and advanced economies) on ASEAN-5’s real GDP. The inverse impacts of policy spillovers implicate the need for
policy coordination at both regional and global levels.

Keywords: Fiscal and Monetary Policies Spillovers, Policy Coordination, ASEAN, Global VAR.

JEL Classifications: C32, E52, E61, R62, F15

After the Asian financial crisis in 1997, economic
integration in Association of South East Asian Nations
(ASEAN) has been continuously enhanced within the
association as well as with the rest of the world, such
as ASEAN+3 and ASEAN+6. Additionally, ASEAN
countries (as a part of East Asia region) have also
become one of the three hubs in the global value chains.
This indicates that ASEAN economies are activating in
an increasingly globalized world with the presence of
deepening regional economic integration. As results,
the member states economy becomes more vulnerable
to external spillovers from global, regional, and intra-

Copyright © 2019 by De La Salle University

Measuring Fiscal and Monetary Policies Spillovers in ASEAN
regional sources which possibly can be transmitted to
member countries’ economy through various channels.
In this milieu, there is a rising concern about internal
(i.e., from any ASEAN country) as well as external
(i.e., the rest of the world) policy spillovers to ASEAN
member economy, as fiscal and monetary policies
adopted in any country in the world can possibly affect
ASEAN countries’ economy.
The effects of policy spillovers could be even more
amplified in time of crisis, as many countries introduce
applicable fiscal and monetary policies to stabilize
their economy. However, without inter-governmental
coordination in policy making, the policy responses
might vary from country to country, which, in turn, may
cause inverse impact on others economies. Especially
for the case of ASEAN, from experience of the recent
the global financial crisis in 2008, many ASEAN
countries and other countries in East Asia introduced
expansionary monetary policy and fiscal policies
in response to the crisis to stabilize their economy
(i.e., stabilize the price and its real GDP). However,
the policy responses were different across ASEAN
countries. These differences may cause an inverse
impact on ASEAN members countries since there are
tightly integrated with each other as well as with the
rest of the world. According to existing literature, small
and open economies (like most of ASEAN countries)
are noticeably dependent on the policy induced by large
countries (regional neighbors and other developed
countries) and have only limited degrees of freedom
in following an independent path (see Capannelli &
Houser , 2009; Auerbach & Gorodnichenko, 2013).
Hence, it is critical to obtain information relating to
monetary and fiscal policy spillovers in a multinational
system. This information is important for ASEAN
countries’ policymakers to monitor their economic
fluctuations comprehensively and then to take in
consideration in national policymaking, as well as
possible policy coordination among deeply integrated
economies, like ASEAN, ASEAN+3, and ASEAN+6.
However, in the literature with the focus on
ASEAN, there is still no study covering this topic
and most of these studies analyze the impacts
of fiscal and monetary spillovers separately by
employing bilateral countries model approaches.
Recently, there is an increasing interest in
applying a global vector autoregression (GVAR)
model proposed by Pesaran, Schuermann, and
31
Weiner (2004) to examine international spillovers
in a multinational system. Hence, to fill this
knowledge gap in this research field, we use the
GVAR approach to comprehensively examine
fiscal and monetary policies spillovers from
both international and intra-regional sources to
the ASEAN-5 economies. This approach allows
us to identify the impacts of such shocks on the
ASEAN-5 countries’ real GDP and inflation
through modeling a multinational system in which
all national and international factors in the system
are interlinked. In the GVAR model, we include
five core variables, namely real GDP, inflation,
real exchange rate, real interest rate, and total
government expenditure, and one global variable,
namely, oil price. Quarterly data of 20 countries
(ASEAN-5, other Asia-Pacific and E.U. countries,
and the U.S.) from 2001 to 2015 are used.
Literature Review
In the last two decades, many researchers have been
interested and investigating international spillovers of
domestic monetary and fiscal policies. However, most
of these studies focused on the policy spillovers from
developed countries (such as the United State, Euro
area, Japan, and so on) to emerging countries or from
large countries to small open countries. Additionally,
spillover effects of fiscal and monetary policies were
examined separately. Therefore, we will review the
literature related to our study in two different strands.
Monetary Spillovers
The first strand of literature is related to monetary
policies spillovers. In this research field, most of
the studies investigated monetary spillovers (both
conventional and unconventional) between countries
by employing different research methodologies—
of which two methodologies are most frequently
employed, namely the event study methodology and
the bilateral countries modeling. On the first strand,
many researchers employ event study methodology
to explain patterns in the unconventional monetary
spillovers from developed countries (mostly from
the U.S.) to other markets (see Fic, 2013; Chen,
Mancini-Griffoli, & Sahay, 2014; Albagli, Ceballos,
Claro, & Romero, 2018). These papers suggest that
32
the quantitative easing of developed countries will
affect the developing countries’ real economy and
the financial market by causing a change in equity
prices, long-term yields, and investment in developing
countries. However, these impacts on the developing
financial market seem to be more significant and larger
than its impacts on the real economy. On the other
strand, there is also a vast number of studies applying
bilateral countries model (such as traditional VAR
approach) to investigate the global output spillovers
from monetary policy of the U.S. and other advanced
countries. The results of these studies suggest that
monetary policy of the U.S. and advanced economies
has significant global spillovers which raise mainly
through spillovers in interest rates (for example, Kim
& Roubini, 2000; Kim, 2001; Faust & Rogers, 2003;
Faust, Rogers, Swanson, & Wright, 2003; Canova,
2005; Nobili & Neri, 2006; Mackowiak, 2007;
Bluedorn & Bowdler, 2011; Miyajima, Mohanty, &
Yetman, 2014). Nonetheless, most of the literature
emphasizes that the magnitudes of spillovers effects
varied across countries, depending on the degree
of trade and financial openness to the developed
countries, as well as economic fundamentals.
Countries with stronger economic fundamentals
(such as higher real GDP growth and stronger
external current account positions, as well as lower
inflation and lower shares of local debt held by
foreigners, more liquid markets) could significantly
mitigate monetary spillover effects.
However, these empirical papers are mostly based
on a two-country model approach, such as two-country
VAR models which involve foreign and domestic
macroeconomic variables of two economies (or vice
versa) and which are estimated for a few countries
only. In a globalized world, two-country models cannot
capture the multilateral nature of global inter-linkages.
The shock to one country’s macroeconomic variables
may affect the rest of the world economies not only
through direct transmission of shock from country to
country but also third-country effects and spillbacks
that a bilateral model fails to capture. Hence, the GVAR
approach has been widely employed by some authors
to investigate the spillover from advanced economies’
monetary policies to others countries such as Chen,
Filardo, He, and Zhu (2015), Georgiadis (2015a and
2015b), and Ganelli and Tawk (2016). The findings of
these studies showed that the monetary spillover effect
of advanced countries to the developing economies
T.M.L. Dau, et al
have been often larger than those to the advanced
economies. Consistent with the studies that employed
other approaches, it is also found that the magnitude
of spillovers depends on the receiving country’s trade
and financial integration, de jure financial openness,
exchange rate regime, financial market development,
labor market rigidities, industry structure, and
participation in global value chains. Furthermore,
economies in its recession phrase will experience larger
spillovers from advanced countries’ monetary policy.
Fiscal Policy Spillover
Regarding fiscal policy spillovers, there are still
limited studies and most of these focused on fiscal
policy spillover of Euro area to other countries
in the region. The first strand of this literature
calibrates macroeconomic models to quantify the
possible spillover effects of fiscal policies (Taylor,
1993; OECD, 2009; Ivanova & Weber, 2011).
They found that an increase in fiscal spending
increases GDP of other members countries.
Another strand of the literature uses a GVAR
model to examine fiscal spillovers in the EU
countries (see Hebous & Zimmermann, 2013;
Ricci-Risquete & Ramajo-Hernández, 2015;
Dragomirescu-Gaina & Philippas, 2015; Caporale
& Girardi, 2011; Nickel & Vansteenkiste, 2013;
Belke & Osowski, 2016). Their results generally
showed that there exist significant spillover effects
of fiscal policy shocks within EMU countries and
these effects are stronger for EMU than non-EMU
countries in Europe. Hence, they suggested a need
for policy coordination among EMU countries.
Additionally, there are some studies which focus
on fiscal spillovers in OECD countries. Auerbach and
Gorodnichenko (2012and 2013) estimated a large
cross-border effect of government spending on output
growth in OECD countries by constructing trade-
weighted fiscal spillovers. Their findings suggested
that cross-country spillovers have an important impact,
and the impact is especially larger when the affected
country is in recession.
In contrast to the study of Auerbach and
Gorodnichenko (2013), Goujard (2013) examined
the output effects of fiscal consolidation in OECD
countries. Fiscal consolidation spillovers are found
to slow domestic growth and decrease employment.
Measuring Fiscal and Monetary Policies Spillovers in ASEAN 33

Spillovers of fiscal consolidations on growth are found area which applies a common currency and monetary
to be initially larger between countries belonging to policy interest rate, they are grouped together and
currency unions. Spillovers of fiscal consolidation are treated as a single economy, while the remaining 12
also found to be more detrimental to domestic growth countries are modeled individually. The time series
during economic downturns in export markets. data for the Euro Area was constructed by cross-section
Despite the vast number of studies on policies weighted averages of all variables over eight Euro-area
spillovers, there is no study focusing on ASEAN countries using the average purchasing power parity
countries and employing multinational countries model GDP weights.
to estimate and analyze the international spillovers in
ASEAN countries comprehensively. Hence, we will The GVAR Model
employ a multinational country approach to fill this In this study, we followed the GVAR approach
knowledge gap. proposed by Pesaran, Schuermann, and Weiner (2004)
Inand
thisDees,
study, diwe Mauro,
followed Pesaran,
the GVARand Smithproposed
approach (2007).byThe Pesaran, Schuermann
Data and Research In this study, we followed the GVAR approach proposed
Methodology GVARbymodel Pesaran, Schuermann,
consists of two andsteps: in the first step,
Weiner (2004) and Dees, di Mauro,
country-specific models Pesaran,estimated,
and Smith (2007). The these
GVAR model consists o
Weiner (2004) and Dees, di Mauro, Pesaran, and Smith (2007). The GVAR model consistsare of two and then
Commented [i14]: There is no corresponding refer
Data steps: in separate country-specific modelsarewill bePls. combined
found. provide. Also, if there are less than 6 author
the first step, country-specific models estimated,
them all here and then these separate co
Given steps: in the first step,
the objectives country-specific
of this study, the models
real GDP, in the
are estimated, and second
then these step. Thecountry-
separate core variables within each
Commented [U15R14]: edited
inflation rate, interest rate, exchange rate, and real specific economy
models will are
be linked
combined through
in the the
second corresponding
step. The core trade-
variables within each econom
specific models will be combined in the second step. The core variables within each economy are
government expenditure are chosen as the main weighted foreign variables.
linked through the corresponding trade-weighted foreign variables.
variables linked
of interest.
through theAdditionally, to account foreign
corresponding trade-weighted for variables. Country-specific models. Assuming that there are
possible common factors, we also include the global n+1 countries models.
Country-specific in the world
Assuming economy
that thereand areonen+1 country
countries in the world eco
Country-specific models. Assuming that there are n+1 countries in the world economy
oil price into the model. is chosen as country 0 representing the reference
and one country is chosen as country 0 representing the reference country—generally, the
The datasets are mostly
and one country is chosen as obtained from the the reference
country 0 representing country—generally,
country—generally, thethemost mostpowerful economy is
International Financial Statistics (IFS) database and powerfulchosen.
economy In our study,
is chosen. In ourthere
study,are 13are
there countries
13 countries included
included (i.e., n is equal
powerful economy is chosen. In our study, there are 13 countries (i.e., nincluded
is equal (i.e.,
to n12
is equal
and to 12 country is 0) and the
one
include the quarterly data from 2001 to 2015.
and one United
country isState
0) andisthe United State
chosen is chosen
0, as country 0,with consistent with the ex
Twentyandcountries
one country from three
is 0) regions,
and the Unitedthat
Stateis,is Asia-
chosen as country 0, consistent with theas country
existing consistent
Pacific, Europe, and America, and one sub-region the existing GVAR literature.
GVAR literature.
GVAR literature.
(ASEAN) are considered in the global VAR models. Using an augmented vector autoregressive model
Table 1 presents countries and regions Using
included inmodel(VARX*) an augmented vector autoregressive
specification, all other model (VARX*)are
n countries specification, all o
Using an augmented vector autoregressive (VARX*) specification, all other n
our study. modeled as small open economies in which a set of
countries are modeled as small open economies in which a set of domestic variables (��� ) is r
As Austria,
countriesBelgium,
are modeled Finland, France,
as small open Germany,
economies domestic
in which a set of domesticvariables
variables ((��� )) isisrelated
related to its lagged values
Italy, Netherlands, and Spain participate in the Euro to its and
lagged a set
values of
and
∗
country-specific
a set of country-specificforeign
foreignvariables,
variables, �∗�� .. Specifically, for
to its lagged values and a set of country-specific foreign variables, ��� . Specifically, for each
country i, the VECMX* for individual economies is represented as following equation
country i, the VECMX* for individual economies is represented as following equation
Table 1 �� ��
Lists of Regions and Sub-Region
�� Used in the
�� GVAR Model
��� � ��� � ��� � � � �� ���� � ��� �∗�� � � ��� �∗�����
��� � ��� � ��� � � � �� ���� � ��� �∗�� � � ��� �∗����� ��� ���
ASIA-Pacific ��� EUROPE ��� AMERICA �
���� �� � ∑���
�
��� ����� � ��� (1
Australia���� �� � ∑����
�
Austria
��� ����� � ��� United States(1)
China Belgium
With � � �� �� � � �; ��� is a �� �� vector; ���∗ is ��∗ �� vector of foreign variable specify to co
With � � �� �� � � �; ��� is a �� �� vector; ���∗ is ��∗ �� vector of foreign variable specify to country
Japan Finland
India i; �� is �� ��� matrix of coefficient∗ for the lagged domestic variables; Λ is �� ���∗ mat
France
i; �� is �� ��� matrix of coefficient for the lagged domestic variables; Λ is �� ��� matrix of
Korea Germany associated with the coefficients of the foreign variables in contemporaneou
coefficients
coefficients associated with the coefficients of the
ASEAN-5: foreign variables in contemporaneous and
Italy
Indonesia lagged form; �
Netherlands � is global oil price; ��� is country specific shock. ��� is assumed to be se
Spain
lagged form; �� is global oil price; ��� is country specific shock. ��� is assumed to be serially
Malaysia UK
uncorrelated with a zero mean and non-singular covariance matrix (��� � �� �� � ��� ∑�� �.
uncorrelated with a zero Philippines
mean and non-singular covariance matrix (��� � �� �� � ��� ∑�� �.
Singapore
Thailand
9
9
 powerful
Weiner (2004) economyand Dees, is chosen. di Mauro, In our Pesaran, study,and there Smith are 13 (2007). countries Theand included
GVAR model (i.e., consists n is equal of two to 12 Commented [i14]: There is no corresponding reference entry
and
nomy one is country
chosen. is 0)
In our and
powerful
study, the there United
economy are State13
is is chosen
chosen.
countries In as country
included
our study, (i.e., 0,nconsistent
there is are equal 13 one with
countries
to 12 country theincluded existing is 0) and
powerful (i.e., the
economy isUnited
npowerful equal State
isfound.chosen.
economy
to 12Pls.is chosen
Inis our Also, study,
chosen. ifas country
Inarethere
our 0,are6consistent
study, 13there countries are with the
13included
countries existing (i.e i
here are 13 countries
en. In our study, there are included
and and 13one
and and
one (i.e.,
country
one
countries one n
country is
country
countryequal
is is
included 0) 0) isandto
is0)
and 12
0)the
and
(i.e., and
the United
n the United
the
is United
equal State
United State
to is
State
12State chosen
is is
chosen
is chosen
chosenas as countryas
country
as country country0, 0,consistent0,
consistent
0, consistent
consistent with with the
with with
the existing
the allexisting
existing
the
provide.
existing and one
there
country
less than authors,
is chosen as country 0 represe
name

In and one
this in study,countrywe followedis 0) and the United
the GVAR approach and
State one
is country
chosen
proposed as is 0)
country
by Pesaran, and 0,the United
consistent Statewith is the chosen existing as country them 0, consistent
here with the existing
steps: the first step, country-specific models are estimated, andGVAR thenSchuermann, these
literature. separate andcountry-
GVAR
ntry is 0)literature.
and the is Unitedand one
State country
isliterature.
chosen is0,the0)as andcountry the United 0, consistent State iswith chosen the existingas country 0, consistent and one country with and the one isexisting0)
country and the
Commented United
is [U15R14]:
0) and the State
edited Unitedis chosen Stateasis country chosen as 0, consistent
country 0
isthechosen
United asState
country 0,GVAR
chosen consistent
GVAR GVAR
asGVAR literature.
country with
literature.
literature. existing
consistent with the existing powerful economy is chosen. In our study, the
Weiner specific GVAR
(2004) and literature.
Dees,will di Mauro, Pesaran, andsecond GVAR
Smith step. literature.
(2007).
models be combined in the TheThe core GVAR variables model within consists
Using each ofeconomy two are Commented [i14]: There is no corresponding reference entry
ture. Using an 34 augmented
GVAR literature. vector autoregressive model (VARX*) specification, all an other augmented
GVAR n literature. GVARvector
found. Pls. autoregressive
literature.
provide. Also, if there are model less than(VARX*)
T.M.L.
6 authors, name
Dau, specification,
et the al UnitedallState other n
Using Using Using ananaugmented
Using an anaugmented
augmented augmented vectorvector
Usingvectorautoregressive
vector autoregressive
autoregressive
an autoregressive
augmented model model
vector model (VARX*)
model (VARX*) (VARX*)
autoregressive (VARX*) specification, specification,
specification, them allall
specification,
model (VARX*)
hereallother all allother
other nother n and
specification,n n one country all other is 0) n and is
steps: inlinked the first Using
through step,the an augmented vector
country-specific
corresponding models
trade-weighted autoregressive
are estimated, foreign modelandvariables. then(VARX*) these separate specification, country- all other n
ged countries
anvector
augmentedare modeled vectoras small
autoregressive
Using open an economies
augmented
model in which a set of domesticall countries
variables are
(��� ) is related modeled as small open
Commented
Using anallaugmented economies
Using [U15R14]:
an augmented in which
edited
vector autoregressive a set of domestic
vector autoregressive variables
model (VARX*) (�
model ) is
�� specific related
(VARX
ressive model (VARX*)
autoregressive countriesspecification,
model countries
countries are
countries (VARX*)
are modeled
aremodeled
are all
modeled
modeled as(VARX*)
other
specification,
assmall asnvector
small
countries open
assmall small
open specification,
autoregressive
all
economies
open
are open other
economies
modeledeconomies
economies n in in model
which inother
which inwhich aopen
which (VARX*)
anset
set aof set
aofdomestic ofofdomestic
setdomestic specification,
domestic variables variables variables
variables(�(� �� )of
other
is)(�
(� related
is)��related
is)nisrelated related GVAR literature.
specific models countries will arebemodeled
Country-specific combined asmodels.
small
in the open
second Assuming economies
step. The
that incorewhich
there a set
variables
are n+1 ofas small
domestic
within
countries each in
economies
variables
economy
the world (�are )inis which
��economy relateda set �� domestic
�� variables (��� ) is related
to itswhich
lagged values Specifically,
and a set offor each country
country-specific foreign i, the variables, �VECMX* ∗ to itsfor
. isSpecifically, lagged VECM
values
forcountries each andmodels a are set(� of )provide
country-specific a complex asforeign
system.
variables, � Hence,
∗
�� .a Specifically, a setfor each
modeled
es
mall in open as asmall
set
economies ofopen countries
domestic
in economies
which a are
variables
set in
modeledwhich
oflagged (�values
domestic ) isaasset small
related of domestic
variables open(�economies )variables
isvalues
related in(�which �� ) foreign
�� a set
related of domestic ∗variables modeled
∗ countries �� is related are as modeled small open economies
small open in which
economies set
in whichof domestic ofvariab
dom
individual to toitsto itslagged
to its itslagged
lagged
economies values �� values and
values and
is aand
to set
and
aitsset
represented aof aset
laggedofcountry-specific
set��of country-specific
country-specific
of country-specific
as following
and a foreign
set foreign
∗of foreignvariables, �
variables, �
country-specific variables, �
evenvariables, � country-specific
�� . ∗�� Specifically,
.∗�� Specifically, . Specifically,
foreign
�� . Specifically, forfor
variables, � each
models for for∗each
each each
. are
Specifically, Using for
estimated an separately.
augmented
each vector autoregre
linked through to its lagged
the values
corresponding and a set
trade-weighted of
and one country is chosen as country 0 representing the reference country—generally, country-specificforeign variables.foreign variables, � �� . Specifically, for
the most each ��
country i, theaVECMX* equation for∗laggedindividual economies is represented as following country equation i, the VECMX* Aas general
to∗�� its forspecification
lagged individual
to values economies afor set the is country-specific
represented
correlation as of following shocks equation ∗
its lagged forand values ofand a set of country-specific foreign variables, �
foreign variable �� . Spe
∗
values
apecific
set offoreign and set
country-specific of to
variables, � country-specific
itsforeign
country
country . country
�� country i, values
Specifically,
variables, �
i, the the foreign
i,VECMX*
i,the and
theVECMX*
VECMX* for ∗ variables, �
VECMX* .aeach
for set
Specifically,
for of
individual
for �� . for
country-specific
individual
individual
for individual Specifically,
economies each
economies
economies economies is for
foreignrepresented
is is each
represented
is variables, �
represented
represented as following as
following
as .following
Specifically,
following equationequation equation
equation each countries are modeled as small open economies
country
Country-specific i, the VECMX* models. for
Assumingindividual ��
that country
economies
there arei, the
isn+1 VECMX*
represented
countries for
asin individual
following
the
included (i.e., across world economies
equation
economy is represented as following equation
powerful economy is chosen. In our �� study, there are 13 countries �� n is equal different
to 12 countries/regions, �� nevertheless, �� can
VECMX*
rmies
and oneand
for
is represented
individual
country
individual
economies ascountry
is chosen
followingeconomies
��� is�isrepresented
�as
i, the
� ���
VECMX*
equation is represented
� as� � following��State
for individual
�����equation
as following
�� � �∗ � � � �∗ ��
economies �equation
� �� ���is � represented as��following
be����country maintained.
���� i, the
equation
��� � ��� � ��� � ���� �� ����
countryVECMX* i, the for VECMX* individual
to �its���
foreconomiesindividual
lagged�∗�� The � values �
is represented
economies
��� � and ∗ a set of country-spec
as followingase
is represented
model. In the nex
one country 0)��country
and the 0United
representing
����� is the ��reference
chosen �� as country country—generally,
�� ����� 0, consistent ∗ ∗ ∗with the
∗ mostthe ∗ existing
∗ ∗ ∗ global VAR
�����
���� ���� ���
� ��� ��� ������������
�� � �� ��
� �
∗� �
���� ���� ����� ����� ��∗��
� �� � �� ������ � �� ���� �
The � �������������� � �����
���
global �
� ����� VAR ∗
model. ∗
In the thethe next �� step, the ��GVAR
�� �� ��� � � ����
�� ���� ��� � � � ��� ���� � �����
�� ��
� �� �� ��
�� ��� � ����� � ��� ����
������ �� ������
�� ��
�� ��
�� ��
� �� � �� � �The ��
�����
The global � �� �VAR
global �� � VAR � model.
��� �model. �� ������ In In next next step, ���
step, the GVAR
the GVAR model ��
model is set is up��up
set by
country i, the VECMX* for
The individual
global VAReconom mo
� powerful
� �GVAR
� �economy
� � � � �is
∗ literature.
� � chosen.
� � �� ∗�In �
our
� The study,
� � �
global there
� � ∗
��
VAR are � � 13�
model. countries
� ∗
� �
In
���������
�
the included
�
���
next
� ���step, The � � (i.e.,
step, � ∗
then �
globalisGVAR�
The equal ������
The model
����
global
VAR
���
to
�
model
�� global
∗���
12 is

isVAR
model. is set
VAR set up
model. upmodel. byby In�stacking
���
stacking
the
�In
� �next �
the � the�
step, � � estimated
estimated �
�
the � �
GVAR � individual
� � �individual
ismodel �
� � � �
upisisby� ∗
country-specific
set � � �
up � by� � ∗
� ∗
stacking�mode �
The global �� VAR �� �� model. In the next the�� GVAR model set In
up the�
by next
stacking �the
step, the next GVARstep, themodel GVAR setmodel issetstacking
set up by the stacki
∗ ∗ ��� ��� ��
� �� �� �� �� �
�� ��� �� � ���
�� �� �� �� �����
���� ��� � ������ �� �� �� �� �� ���� �� ��� ����� ��� �� �� �
� ���� �����
���������
�� ��
�� ����� ��
The global �����
VAR model. ∑In the ��next step, the GVAR model up by (1)
�� ����� �
� ∑��� ����� ���
� ��� ���
∑� �� �� �� ��
∑ �∑ �∑
�� (1) The ����
The
�global
(1)
estimated global
estimated
�country-specific VAR The VAR ��
individual individual
model. ��
global
�
model. � country-specific
models
InVAR the In���
(1)(1) nextthe
country-specific and
model.
(1)
��
next
(1) step,linking���
Instep,
the
�models
modelstheGVAR
�the
���and
them
��
next GVAR and linking
with
model
step, model
linking
thea ismatrix
��� them
GVARsetthem isup setmodel
with by
���upastacking
with by aisstacking
matrix stackin
matrix
set
�� ��� of
the
up
and one country ��� ���
is 0) an and the United ���
State is ����
chosen
�� �� �� �
as�������������
country
� � �� 0,
����
�� �� � ��
consistent
�� ����
�� ��
����
��� ���� � ��
��with
��� �� �� ��� �
� the �� ∑ �� �
existing � �� � � (1) estimated individual country-
Using augmented
estimated
estimated �� vector
individual �
individual � ∑
autoregressive �
country-specific
country-specific ��
model�
��� �
���
models
���
(VARX*)
models estimated
and and linking specification,
linking
estimated estimated
individual them them
individual
��� individual all
�� other
with
country-specific
with a a
matrix(1)
��� matrix
country-specific n
country-specific
of ��of predetermined cross-country linkages.
models
predetermined models
and modelslinking and and linking
them linking with them them
a with
matrix with a ofmatrix
a matrix
predetermined of predetermi
of predeter
�� � ��� �� ��� ��
With estimated of
individual predetermined ��� iscountry-specificalinkages. �� vector; cross-country �models
∗ ∗ and linkages.
vectorlinking � of foreign them �
with �
�a��
�matrix � �
�of
� �
���apredeterm
� � ��
a variable
∗ ∗ � � ��
estimated ��to � � �; country-specific
individual ��linkages.
country-specific models isand � �� and �
∑linking them � �with a��them �� specify
matrix of to country
predeterm
With � � ������
�� ��
��� � ��� �;� ��
��� is∑���� a� �� � �� vector; � �
�� ����� � ��
is
(1) ��� � �� vector of
�� �� � ∑���(1)
�� foreign variable
���∗ �� ∗ �(1)
∗∗∗estimated ∗�∗��
specify individual country
cross-country
cross-country
estimated individual (1) models
country-specific ���� ����linking � � models �� them�����and ���with ∑���
linking
��
�matrix ofwith
��
predetermined matrix o
�� �� � ∑���
���
GVAR countries
�� literature.
��� � With
��
With � �are��modeled
�� ��� t
With =
With �With
�� �cross-country
� �;
� �1,2,
With
as
� �
�� �
���small
��� �
��
�...,
is a �open
����
�� T
�
��
����; ;
��
���; ���
� �
vector;
��linkages.
�;
economies
�
� is
�;is
�� ��With � a
is �a� a
is ��
is
� a
�����∗��� is�in
��a
�vector;
� �∗���
�which
� ��vector;
�� vector;
vector;
� �� � ���
vector
vector;
�
aglobal
set
� ∗ ���
is � �
is �
�;��of���domestic ��
is
�
� ��
foreign
is is�
�� vector
� �� vector
is� a�cross-country ��
���variable
�� vector
variables
of
vector
vector; offoreign of
foreign
of
specify
foreign
∗ foreign
Due
�linkages.
(� isvariable variable∗
��related
to
)linkages.
is
variable
to
�� vector
country
specify
variable specify specify
the contemporaneous to
specify country
to to
country
of foreign variable specify to country
country ��� ��
dependence Due to country cross-country
���
to the contemporaneous
��
of the
���
linkages. ���
depende
cross-country �linkages. � ��The cross-country
��
VAR model. cross-country
linkages. In the ��
next
�� step, the GVAR model is set up by stacking the
�∗�vector of� ��foreign ∗ forvariable
�specify to �variables;
country i; isis
i;vector cross-country linkages. ofof�Due coefficient for the ∗
�is �cross-country ��foreign
�� linkages. �alagged domestic variables; ∗ Λ ∗is ��� �matrix ono
∗ matrix �variable
�� i;���∗��;
a ������
is is �vector
�� ���� ��
is �a matrix
�of �� ∗of coefficient
�vector;
With
foreign variable
∗
isof
�� ��� ��
��specify
�� � �;vectorthe lagged
�variable
to isofa foreign
��country �� domestic
vector;
variable ∗
specify
is Λ
��∗ ��cross-country
to country �of � �� � �matrix domestic
linkages. With
cross-country
variable Due � variable
tospecify
��the
to
With
�� thecontemporaneous
� �to
linkages. �∗contemporaneous
�; � ∗country�∗����
�� ∗ is on ���;foreign
��dependence dependence
���vector;
is a� variables
����
∗ of
�� is
the
vector;
� of�∗ ��thedomestic , the
domestic
�vector
�� is �of variable
� �� foreign
�vector
∑�� ��on
variable
�of forei for
∗ � �� ∗ matrix of Due to the contempora ��
vector;
Using an is i; i;�
� ��
�� augmented vector
��i;
is�i; ���
is
��
vector ��
�� is is foreign
� ��� matrix
� ���� � matrix
� matrix
� ��
autoregressive of ofcoefficient
� matrix i; of specify
coefficient
coefficient
�of
model coefficient to for
for
(VARX*)country
the
��for the
for lagged
the lagged
the
of lagged domestic
lagged
coefficient
specification, domestic domestic
domestic allforvariables;
variables;
the
other variables;
variables;
lagged n Λ Λisdomestic Λis�Λ� �� is
��is�� � matrix
��� ��matrix
variables; matrix of of� Λofof
matrix is �� �� � �� ��� �
toi;its��lagged is �� ��values matrix
� matrix andofDue aof coefficient
coefficient
setto
Due ofto country-specific
the the for
contemporaneous the
contemporaneous for
� is
� lagged
estimated the
� �� lagged
matrix
domestic
� foreign� dependence
dependence
individual domestic
variables, � variables;
of of Due
the
country-specific
∗
the .toSpecifically,
domestic domestic
Λthe Due iscontemporaneous
Due �
to � ��
variable variable
country-specific
the
to
∗
�the matrix
for
contemporaneous each
� on
�contemporaneous ofon � �
foreign
dependence
foreign
� �
VAR
� �
variables variables
dependence models
dependence
of the �
ofof��the
domestic ,
need
of the country-specific
ofdomestic
� the to be
� domestic
variable variable solved
��variable
�variable on foreign VAR
���� �on models
foreign
on
��onvariables foreign
��� need
variab vat
∗models and �� �� linking them with a the matrix predetermined
��
coefficients associated with the coefficients of the foreign variables The ∗ coefficients
in The
contemporaneous
global global VAR Due Due associated
VAR
� model.to to
, ∗the
the the
model.
and contemporaneous
In with
contemporaneous
country-specificthe In the
nextthe ∗ coefficients
next
step, VAR dependence
dependence
step,
the GVAR of
the
models GVAR of
model
need the
foreignthe domestic
model
to is
be variables
domestic
set
solved is up set variable
by up in by
stacking
simultaneously contemporaneous
�stacking on
the foreign
foreign
for the all varia
dom and
vari
Due to the � i; , the
contemporaneous
� is Due country-specific
� to
�� i; the
� matrix is dependence
contemporaneous
� ��of VAR coefficient
matrix models
of of
the need
for
coefficient
domestic
dependence the to be
lagged solved
for
variable
of the thedomesticsimultaneously
domestic lagged on ���� variables;
domestic
foreign
variable for
variables Λall
∗ d
varia
on is
thevariables;
offorcoefficient i; coefficients
for isthe is� matrix
lagged domesticmatrix
of coefficient isof
variables; coefficients
for
∗ Λ the is lagged associated domestic
matrix offoreign
variables; � isin��contemporaneous � �� matrix
all�of
matrix domestic
�lagged variables; � �� �� �associated
� ��
∗
matrix of �of �� �� �� Λ � �� � �
issimultaneously for domestic �variables (for
∗
�� � � � �� ,�the country-specific �VAR
oefficient lagged �domestic
coefficients variables; �associated with the
with matrix �
acoefficients
the coefficients � of �the of foreign
the theforeign variables variables in
coefficients coefficients associated associated with withthe �coefficients
the coefficients of theof foreign )variables variables ofincontemporaneous contemporaneous
inforeigncontemporaneous and and and and
With
Λ coefficients � �� a domestic
�� �� vector; ��
countries are modeled as small ∗ � ∗ open economies in which � �� set associated
of domestic with
variables the
∗ in∗coefficients
(� related the variables in contemporaneous
� � (for ��i=1,2,…n). �;and ��The is and is
fore
coefficients
country associated
i, the VECMX* ��� , �� , with
for
the
the country-specific the
individual
country-specific coefficients
economies VAR of isthe
models foreign
represented need variables
∗toas
���be, the befollowing
solved
� ,
country-specific the contemporaneous
, theequation
simultaneously
country-specific
�� country-specific VAR formodels for and
VAR all VAR domestic
models
need models to variables
need
be need
solved to beto solved
be solved
simultaneously simultaneously
simultaneously for all domestic for all
for domestic
all domestic
variables
��
variab va
with the oilcoefficients of VAR the cross-countrymodels need
foreign linkages.to
variables solved
lagged � ���in ∗simultaneously
∗ �
����,form;
��the
to country-specific i=1,2,…n). is
allVAR domestic
The models
domestic
variables need to
and be
��
solved
foreign simultaneously
variables for
are all domestic varia
lagged
sassociated form; with��the
ofthethecoefficients
foreign is global
coefficients
variablescoefficients in price;
ofassociated
contemporaneous the ��� is foreign country
with variables
and the specificcoefficients
in shock. �
estimated
contemporaneous estimated
of is�� the isindividual
� ∗ individual
assumed
�� ,specific
foreign the and ,country-specific
the
variables be
country-specific ��
country-specific
country-specific
�� serially
� ��(for ∗ (for
��in
�� , global
coefficients the i=1,2,…n). i=1,2,…n).
VAR oil
models
country-specific
contemporaneous VAR price;
models
associated
coefficients
models and
The models The�need
��and is
linking
domestic with
associated
VAR
and country
need
domestic
to linking
be them
the to
models and
solvedbe specific
themsolved
with
foreign
coefficients
with
and need awith
the
foreign
simultaneously
to shock. �
simultaneously
matrix aofvariables
variables matrix
coefficients
be the
solved is
of�� predetermined
areofassumed
foreign
for
(for
of
are
all forthe
grouped
simultaneously
i=1,2,…n).
allforeign
predetermined
variables
grouped
domestic toas a
domestic bevariables
as a
for
The
inserially
(� vari
�do
�cont
variab
all(�
dom
�
th
to its lagged values of the
lagged
contemporaneous
and foreign
lagged lagged form;
lagged form; variables
form;
form;
� �andis � is� in
global is
global
laggediscontemporaneous
global
laggedoil
global oilprice;
oil
price;
form; oil
form; price;
price;
� � is �
isand
country
is
is � country
is country
global
global∗ specific
country oil oil specific
price; shock. �
specific shock. � shock. �shock. �
country is isassumed is
assumedis assumed
specificassumed to to be
shock. �to
be serially
to beserially
be serially
serially
is i;
assumed � is � to �� be matrix �
serially of coefficient for the la�
lagged form; ���ais�set global of
(for country-specific
oil �
i=1,2,…n). �
price; � �
� �
The� foreign
is country
domestic variables, �
���
specific
and �� �� ��
foreign
�
shock. � �. Specifically,
variables is assumed are � for is
grouped
grouped each
��
to ��
be ��
as
��
serially
as a a (� � � ∗
��� vector,
vector, � �
� �
�
� ��
� �
to to rewrite
��
rewrite the the country-specific ∗ ∗ V
���∗���(for i=1,2,…n). The (�domestic � and foreign �variables ���are grouped (�� as a �(� vec
� �i=1,2,…n). �� ∗ �� ∗ ���vector,
��∗���
�� (for �� i=1,2,…n). The �� domesticDue andtoforeign �� (for
��
the�contemporaneous variables ��
(for (for
are i=1,2,…n).
grouped The �domestic
dependence
i=1,2,…n). as a ofThe � and
the � domestic
��foreign
�domestic ��� and
vector, variables
variable foreign are
�� variables
on groupedforeign∗ are as a
variablesgrouped � as a
� ��� � (� � � ���
�� �The domestic and �foreign variables are grouped ��as a (� �������as a ve
�� �� ∗���
uncorrelated
is specific
�� price;
ountry
oil global �
with price;
a zero
oilshock. �
is price;
country lagged �mean
��is �isform;
is
specific country
and
� �non-singular
country
assumed
�� ���to
shock. �
�� is �specific
specific � serially
global
be
�� is
�� oilshock.
covariance
shock. �
assumed
�price;
� ���� ��tois � �cross-country
�matrix is
��assumed
be is ∗(�
�country
serially
��
cross-country
assumed
�� � �� to � be
� �uncorrelated
�� � ��linkages.
��to
specific ����
� ���
��serially
(for belinkages.
�����
(for
∑ �� i=1,2,…n).
shock. �
i=1,2,…n). with �.���
�� ���� a��
lagged
�country-specific��is
zero
The
(for ��
∗assumed
mean
The
form;
to
� rewrite
domestic
∗i=1,2,…n).
��
to and
domestic
laggedto��be
rewrite non-singular
isthe
VECM
and form; global
serially
The the and
country-specific
foreign foreign
�oil
country-specific
domestic models covariance
price;
isvariables
global and
coefficients
variables
(1): oil is
�VECM
�� are
foreign matrix
VECM are
country
price; groupedmodels �(�
grouped
variables
associated models
����
�
specific
is as a ���are
country
(1):
with (1):as a
� ���
��(�shock. � �(�
�grouped
the ��∑
specific
�∗��
� ���
coefficients
��. ���
is vector (�ve
assum
shock. �o
country i,uncorrelated
the VECMX*
�� uncorrelated
for uncorrelated
��uncorrelated uncorrelated
individual with with a
with
��� economiesuncorrelated
��zero
with
a a
zero mean
zero
a zero
mean
is and
mean mean
representedand non-singular
and non-singular
non-singular
andwith non-singular
as following
a zero covariance
mean covariance
covariance covariance
equation and matrix matrix
non-singular matrix (�
matrix (�
�� ��
�� (��� �(�
�� � �� �
��
covariance
�� ��
�� � ���
��
� ��� ���
�� ∑
� ���
�� ∑
� ���
�� matrix
�� ∑ �.∑
�� �.
�� (� �. �. � �� �� � ��� ∑ �. � �� � � ∗ � to rewrite the cou
serially with a�� zero
���uncorrelated
�� �����∗�mean
�
� ∗ � to and to with
rewrite non-singular
rewrite a���zero
the the ,mean
the
country-specific covariance
���country-specific
∗
and non-singular
country-specific matrix
VECM���VECM
��� � �(�
�
VAR
models
��
� �
�∗�����models�� ����
tomodels
�
���∗�� ���
(1): ��
rewrite
�
���∗(1):
� ∗ to
��
need
� ∑ rewrite
theto �� rewrite to��. bethe
country-specific solved country-specific
the simultaneously
country-specific VECM �� VECM
models for all
VECM models
(1): domestic
models �� (1):(1): variables
� � �
� �� ∗
� �
�
∑��� ��� ��,��� �
��� Due Due
to the to � the �
��contemporaneous
�
contemporaneous to rewrite the country-specific
dependence of the VECM domestic models variable (1): on foreign variables
�����∗dependence of the∗with domestic variable on foreign(1): variables
�� ���(1): � � ��
�� ��� � � to �
∗ � �� to rewrite the country-specific VECM models ��
� � ��
�
with a zero mean anduncorrelated
non-singular ��
with
covariance a(��zero � mean matrixand (� ∑non-singular
� ���. �� .� ��� covariance
∑ �� ����.� ��matrix ∗ � ����
�� ��
�� (� ���uncorrelated
rewrite �����the � ���� to ∑
country-specificwith
uncorrelated
rewrite ���a�.�� zero
∗the mean
∑� ∑a����
VECM
country-specific �and
zero models non-singular
mean �and
(1):
VECM covariance
����non-singular
models
�� matrix
covariance (���� matrix
� �� ��cou � (
� is global(3)
lar covariance
ean and non-singular matrix (�
covariance �� � �� ��matrix
covariance � ���
matrix ∑�� � �. ���� �� � ��� �� �� �� � � �� ��� �
�� �� �� �� ���� �� �
lagged��,���
�� ��,��� �form; ��� oil price; �� is ∗
���� �� ��� ∑��� � �� �� � � ��
(1) ��
∗ � ��� � ��� �
�� � ��� ��� ∗ ���� ∑
��∗
����
(for
� ���
∑ �
�i=1,2,…n).
��� ��
∗� � ∗The domestic and�foreign variables
� � are � grouped (3) ���as a �where (�� ������ ��� ∗
� vector,
�domestic �� �� �variables� � ∑variables
� �
��� ����� (;
�,� ��� country-specific
∗ �
�In the ���GVAR methodology, the vector �of country- � ∗� � �need ∗ ∑ �∑ (3)
�� � � ��� � � �
� � ��� ����
�� � �� �∗��
��� �
�� � �� , �,���
�� the
� ∗�
��
the
��country-specific �� �VAR �models
VAR�� ���� � �models∑ � ∗∗����� ��
need
to ∑ be� ��� ��
to �
solved ��be�(3)
� ��,��� ��
solved simultaneously � �� �� for
simultaneously allfor �� all domestic ���(3) (3)
���∗������ ∗� ��� � ∗�∑ ��� ����
�� ��� �� ��� ��� �,��� �,���
�� ��� �,��� �� �����
where
where� where
� �
� �� ∗�� � � ���∑���� �� � �
�∑���
� ���
��
��� � � ������� ����
�,���
��,���� ��
∑ ∑� � ����
uncorrelated
�
��� ; � ;
� ; � � with ��
� a
��, � Λ
zero , � Λ �;
mean � �; ; �and ��
� , Λ
non-singular
�� ,(3)
Λ�; the
�;(3t
specific
�� � � �;foreign ��� is a ∗�variables, vector;9����, isplays ∗ ∗
avector
central role. As a ∗ specify � �� �� �� �� � ���� �� � �,��� �� ��� �� 9�� �� ��� �,��� � �� ��
where ��� � �� ∗ � � ��
∗ ��� � ��� �� � ∑
With � � �� ����� �� � �� ��� of foreign variable to country �� ��
��� �� � ���
��� �� ��� � �� �� � � ��
��
�����to rewrite 9the ∗country-specific VECM
�� �∑�� ��, Λmodels �� (1):
���∗���
∗ � �� � � � � � ��
where � �� (for 9; �
(for
�i=1,2,…n).
;where �9where 9��,�
i=1,2,…n). where ,�The
� Λ������� ∗�;�� �
The
domestic
� ��domestic �� �� ∑and �� ; the
and
foreign
�� �;
�∑
thedimensions
foreign
∑ �dimensions
variables
� ��;∗��� ���variables �� ; ;��
are of
�of ,;�� Λ �� ��
�are
grouped and and
grouped
�;��, ,� ��, �are
as a are �;
(�
�� as a
���
���;
�,x x(� �(� ∗��
� ��;
�
the
���,�,Λ��∗��dimensions
,�, and
Λ�;�;
vector, ��the�;vector,
the � has
�dimension of a ful
dimensi
way ofwhere the
��� ����∑ � � �� �;
∗∗� ���� ��; �dimensions of � ��
��� ���
� � �� �� � � �
�� � �� � ��� � �� �� Λ � Λ
��� �� ��
�
��� ��� ��� � �� ��� � �� � � � Λ � 9� ,��Λ���
dealing
��
with ∑ �� the curse of dimensionality
��� 9��� ��� ��� ��
when
where
��
�� where
where
��
n ∗ ���
�variables,
��
�
��
��
��� ��
�where
�(1) and
�
� ��
������ ∗�and
�
��
�����are �∗����
� �
�����
� �
�∑
are
��
�� x�
��
��� ∑
����∗�(��
���
∑
x,���
����
� � ���
(���� � ����
�� ��
�����, and
� �∗
���
;���
�, and∑����
�
�� � ;� �
� has ����� � �
has
��
��
��
a, � Λ
��
��
full��
��
a��full,column
� Λ
; �; ��
�column
����
�; �;
���
� ���
rank �
��
�� �ki . ���;
rank
, � Λ
����
, Λ ���� ki9,�;
Λ ��
.��the
� �;� the
the
dimensions
dimension
��dimensio
, Λ �; of th
�� �coefficient
�� � �� � � �� �� ��� ��
9 i; � � In
is �the
� �� GVAR
matrix
is relatively
�9 In methodology,
the
of GVAR
large, 9 ���
the the
methodology,
for
��
weights vectorthe��� of
lagged
∗ � ∗used
thecountry-specific
�� vector
domestic
in the of 9
construction
�full
foreign
country-specific
variables; ���∗
Λ ��
� is foreign � (and
��
� �� ���� ∗ ,�� has
plays
variables,
matrix � �� � a afull
of �
���
∗ ���
column
����
plays
�� �
�� ���� a
� � rank � k i
�.
�� �� ��� ��
Then,
� � 9 �� and �� are �� x (�� � ��
��
gathering
�
the
�
variables
�
from
��∗ �, a
al
� and � are � x (� � �, and � has � a � �� �column � � ∑
and rank � k are .� x � (� � ∗ has a full column rank k . (3)
�� and � �� are � �� x�(�� �� �� �, and �� has �� �� a�full ∗��
��������� and
column
� �to � �
rewrite
� ���rank to
are ��rewrite and
� �
k�the
ix �. � (� � �the
are ��� country-specific
�
country-specific ���� �, and x��� �
(��� � � �, and
�VECM
has� �, and aVECM
� full� models
��column has amodelsa(1): full rank (1):
column ki . rank rank kiki i.kcountries
i.
i
� � of are not estimated
� ��� ∗��
� and
�specify and ���
are�
� ��
�� are
to�are
�,���
Then,
� �� xThen, x(� (� gathering
� ����∗��gathering
∗�, and
���, and
gathering � �
the the
�� has has ∗variables
variables full
full�column
acolumn column
from fromall rank all
countries . inrank ainglobal in vector
i . vector thethe system
the�but ��specified a priori, �based
� and x �n(� and �Then,
�����∗theare x��(�has athe full variables � has from
rank a full kall countries
i . column akglobal sys
��
With �central � �� ��role.
coefficients �;As �a�� is
central way
associated a �of
role. � �� As
with vector;
dealing a way thewith �of��∗ is
dealing
coefficients
∗
�curse vector
withofofthe theofforeign foreign
dimensionality
curse variable
ofvariables when
dimensionality nin� � iscontemporaneous
� relativelywhen �
�country � �
large,
is relatively �, and
and large, � � the � � �, and Then, gathering �the va
on information Then, Then,
that gathering
measures
gathering thewherethe variables
the �strength
variables ∗
�� � from from
���all all
�ofcountries countries
�bilateral
�� ��Then, � ∑ in
�� in a global
Then,
gathering
a global
� ∗ �� a
Then, global
gathering
vector
∗ the vector
�gathering; variables
� vector
the
�� system the �� thesystem
variables
, fromthe
variables
� Λ � system
all�
�; � from
countries
� from� all
�� countries
all , in
Λ countries
a �;globalthe in ain global
vector
dimensions a global thevector vector
system
of the system
the
� � system
�� � ��� ���� , ���
∗�
�� ������ ��� �� Then, ��
Then, ���∑gathering���
gathering�� ∑����� � ��,��� the
�the �� ��
�� variables
���
variables � ���from
��
�
from�
�variables allall � countries
� � ∗� in
��
countries in
∗� a a global
global vector
vector the
(3)
the � (3)
system
system � �
�
i; �� isweights �� ��� matrix used in�of�the
weights coefficient
is used construction inoil forprice;
the the lagged
of economy.
construction ������ isare domestic
not
of estimated variables; butestimated Λ
specified is �Then, ∗
��a� priori,�� �� gathering
matrix ��
to based
���
ofaserially the
Then, on variables �,���
gathering from theall ���� ��
��countries � �� , �in
��from ����, a��
�all ���countries
global � �� �� � � �� inthe
vector a global systemvector �� the �� sys
lagged form;linkages global
in the global country �� are not
�specific shock. � �� is assumed � but specified be priori, based on ��
�� and ����� �� ���
� �are ��� � ��∗�
where �, �
�� � where
,
x ���
(�
∗�
�� ��
�
��� ∗ ���� � � � � �
� ∗� �∗ � ��
� �, and � has a � full column
� � ��� � ��rank
;�
�� ��∗���
, �
k
��������;��
�� � . ���
i�� �� � ��
� (4)
(4) ��
��
�� , �� ∗� ��∗�� where
�
,����� ,∗��
��� ∗��� ��� ��
� ��� � ��
����
� ��� is a (ki +ki*)x(4)
� �� ���;, Λ �; the dimensions of
the dimensions
k matrix (4) of fixed(4)
of terms
kno
(4) (
In this paper, we use trade weights. The choice ���� � ����where � � ∑ ��� ∑ �� �(ki �� ��
�� �,matrix
� Λ , � Λ�; � �;
� �� ,��Λconstants (4
�
coefficients associated with thethat coefficients ofbilateral
the foreign variables �� in (� contemporaneous
�� �� �� �� �� where � and�. � ��is
�
��� a is a
��
��� � +ki*)x
(ki �+ki*)x
��� �� k� ���
matrix
,
��k � ∗�
�� ���� , �
�of
��
� �� ��
fixed
�of
� �fixed
�
����
� ��known
, �known
�� ∗�
� �
�� constants
� � �
�� defined defined in terms
in of (4) the
of co
the
information
uncorrelated thatwithmeasures
information a zero the
mean strength
measures
and non-singular ofthe strength linkages
of bilateral
covariance inmatrix
the global
linkages ��
economy.
in
� the
�� �� global
� ��� ∑
���
economy.
��
� �� �� �� ��
�� � � �� ��
� where
� � � � is a (ki +ki*)x k mat
of trade weights
where ��� is� ais(ki ais(ki based
+ki*)x kon matrix two of considerations.
fixed known constants �definedis∗iswhere in is
terms aof�k(ki of +ki*)x
the incountry-specific k matrix ofthe weights
fixed �defined
known . It can constants be viewed asthe “link” matrix
where where iswhere a (ki � +ki*)x �aisaa(ki a∗kterms+ki*)x
(ki matrix +ki*)x of matrix
kfixedmatrix aofknown fixed
of fixed known
constants known constants constants
defined � defined
in�defined in terms
in terms ofofthe of country-spec
country-s
where +ki*)x k matrix ofThen, fixed known
gathering constants
�
the variables defined � from in all countries the country-specific global vector system �� terms of the country-specific
� and � � and are � � �where
are xwhere (�� x
� � � (� �
��weightsis�
�, and (ki
� (ki +ki*)x
�, and
�+ki*)x
� �has�
. It kcan
a k matrix
matrix
has
full be a full
column
viewedof of fixed
fixed
column rank
as known
known
k
“link”rank . constants
k
matrix .
constants that defined
allows in
thein terms
terms
country-specific the country-spe
ofcountry-specific
the country-spe model
lagged form; In �� this is global
paper, In
First, trade in goods oil we
this price;usepaper, �trade is we country
weights.
use specific
trade
represents an important channel The shock. �
choice
weights. � of
The � � is
trade where
assumed
choice � weights
� of � � is
to
trade isa be (ki
�� weights
+ki*)x
where
serially
based
weights � on � � k is� matrix
two is .
a
based
� It
(ki can of+ki*)x
on be
fixed viewed
two k knownmatrix as i “link”
constants
of fixed i matrix defined
known that allows
in
constants terms the of
defined country-specific
the in terms of the mo
��
�� � . It can be viewed as “link” matrix that weights
�� � defined ��
in terms of the country-specific weights
� �� weights ��� . It can be viewed �
weights allows �
��weights
the���� �country-specific
.�.It �
.can It
���can be viewed
be viewed models
asmatrix “link”
as to
“link” be matrix written
matrix that in
thatallowsterms
allows theof global
country-specific
the variable
country-specific vector
models models � �to be wri
weights . ��
It can be viewed as “link”across matrixweights that allows .�the
It��can be
country-specific viewed as “link”
models to be that written allows the country-specific (4) models to be written to to be w
��� � � � ∗�
through which shocks arein transmitted countries, � �� It , can
can ��be � � � �all
ofbe �viewed as “link” �matrix that allows � the
�
Then, weights gathering �� It the variables viewed � as “link” matrix that allows the country-specific �� models be wr
Then, ��gathering
weights � the .weights
It
variables can be fromviewed asfrom all
asvariable
countries “link” countries matrix
in a �globalin thata global allows
vector the vector the system
thatcountry-specific
the system models to beto bemo w
�in �� �the
uncorrelated considerations. zeroFirst,
with aconsiderations. mean trade andinFirst, goods
non-singular trade represents goodsanrepresents
covariance important
matrix (� an channel � ��weights
�� important through
� ��� �
channel
∑���� .which It ��
�.can
�� terms
in
through be terms
shocks viewed � global
of
which �� .global
are It canvariable
“link”
shocks be matrix
viewed
are vector vector that
as � allows
“link” � matrix country-specific allows models
country-specific written
especially for theglobal case of where ASEAN, in terms of global variable ve
in termsin terms of of global variable variable vector vector
9 � �� � is� a trade(ki +ki*)x in terms
linkage k in matrix
of
in terms
in
global
terms terms ofofof fixed country-specific
global
of
variableglobal known
global variable
vector
variable constants
variable�� � vector
∗�� vector
�vector
models ��� �� to
defined
∗� in terms be written inCombining
of the country-specific terms of equations global (3) and (4), we
transmitted is much stronger than for financial linkages. Second, in terms of global � � ���
variable
Combining , ��
� ��
vector ,� �equations
equations � ��� ������ � (3) �(3)
and (4), we get (4) (4)
across
transmitted countries, across especially
countries, the case for
especially of ASEAN,
the case trade ofinASEAN, linkage
terms of global trade is variable
much linkage
invariable terms of
�� stronger ��vector
Combining
is ��much
vector
global�variable
�� �� stronger
�
�� vector ��
� � and (4), we get
Combining equations ∗ �
trade data isCombining Combining
more available
equationsequations weights (3)
in(3)comparison
and �and� (4), we get
. It can
(4), we beto
get viewed other
Combining as “link” Combining
Combining Combining
matrix
equations Combining equations
that equations
(3)
equations allows
and (4), (3)
equations
the
(3) and
(3) weand and (4),
country-specific
get (4), (4), we (3)
we get
we ∗get andget models (4), wetermsto get be written ��� �� �� � ��� � ∑���
than financialthan linkages.
financial Second, linkages. tradeSecond, data is trade more data where
��
available is more where
in is �
�� comparison a (ki
available � is+ki*)x a (ki in to+ki*)x k
comparison
Combining equations matrix
Combining
other k matrix of
alternative to fixed
equations of
other fixed
known known
(3)
alternative
(3) andequations � constants
and
(4), � constants
�(4), � defined
we � get � defined
∗∑ in � terms
∑we � in
� � of � the of� the
country-specific
� country-specific
Combining ����� �we �� �get �(3)� � and ��� (4), ���� ��get � � � ���� � �� ���
alternative measures that could ∗be∗ used � because � ∗ ∗ (5) � �
�� ���
By
�� ���
stacking individual models:
��� �� �
�9����in �terms � �of�� ��∑��� ����� ����� �vector ������� � � ∗� � � � ∑ � � � � � (5)
��� � �� �� ��� global∑ variable
� � � � viewed � � � � � � � � � ∑ � �� (5) �� �∑ � � � � � � � (5) (5) (5)(
�� � ��� ∗ �
measures thatmeasures reliable
could bethat bilateral
usedcould because be used trade
reliable because statistics
bilateral reliable are
weights
tradebilateral ���
statistics published
weights � ���
.
trade
�� It� can�
are published
�� . Itbe
statistics are can viewed be annually as
published
�� “link”
By as
�
�for By
��
� �
“link”
�all
stacking ���
matrix
annually ����
��
� ��
matrix
�
���� �� individual
�stacking
�� � that
� �
����for
� ∗
�
� ��� � �
��all
that
allows
��
individual
��∑
∗
��
���
��
�
�����
� ∑
allows
��models: ∑ the
�
��� �
���
������
���
���
��models:
�
� ���
��������
the
�� ∗ �
country-specific��
��
� �
�����∑����
���
country-specific
��
� �
���
���
�
� ��� � ��� �
��
� � ���
��� models
��
�� � ��� � �By
models to be to
written be
�� stacking individua
written
(5) (5
annually forByall By countries
stacking
stacking individual
individual by reliable models: models:
Combining international equations By stacking (3) and By Bystacking
(4), stacking
individual we get individual individual
models: models: models: ∗ models: � � �� �� � ��∗ � �� � � ∑
countries bycountries reliable internationalby reliable organizations.organizations.
international Datainon terms in terms
bilateral of global
Data offinancial
on global
variable
bilateral By
variable
By flows stacking
vector
stacking
financial By are vector�stackingindividual
�eitherindividual
flows �stacking are individual
models:
models:
� either � � � � ∗� � � ∑ ∑ � � � �
organizations. Data on bilateral financial flows �
are By stacking individual By � models: individual �
� �� � � � � � � � models:
� � � � � ��� ��� � ���� �
� ��� � � � �
∗ �� �� �
�� ��to � � ��∗much���∗ ��� ���� �∑ ��∑��� � ���� � ����� ∗ ∑ �
���∗� �� � �� ∗ ∗�
�� ∑ � � �
(6) � �(6) � ∑� � �
�∑ � � � � � �� � �(5) �� �� �(6) (6) (
nonexistent or either nonexistent
tend to be or
nonexistent much tendmore or
to bevolatile tend
much more �� ��be
and less volatile reliable, more
and less asCombining
��� �volatile
their
�� � ��
reliable,
��
Combining
��� �
collection
�
� � � �
equations
�
as their equations
has
�
collection
�
(3)
started �
��� �
and
� � �
(3) �� (4),
only and
has
� �
�
���we
� �
���
�
�
(4),�
started
�
�
�
�
� �
get � �
�
we
�
� ∗��
∗ � �
only
�get �
����
� � �
� � �
���
�
� ∗∑
���
���
∑ �
∗ �
���� �
�
� � �
where:
���
� � ��� �
� �� ���
�
� � �
� … � �; � (6)
� ∗
� � …
��
� �
� � �� �
(6)
�(6
�� �� �������� � � ∗ ��
� �
��� ���
�
� �∑
�� � �� ��� ∗
���
�����������
� ���� � ���
� ���∑��
��� �
� � �
�� � � � �� � �
��� � �� � (6) � ��
and less reliable, as �����
their
�����
collection ∗ � ∗ has started � onlymodels: �� where:
����� �������
where: �∗�� ������ ∗ ���� … …�; � ∗ �; � ∗
∗� �� ��…� �
∗
� ���
�…�� � � � �� ���
� � ��…� �
���
�…� �
�
�� �
���
� � �� ��� ���� �… …
�
�
� � � � �� ∗
� ���… ��� ��…��∗…� �
�
more recently. more recently.
��� By �� stacking�individual �� �� ����
���
� � �
�� � � �� ��
∑ ∗ � ∑����� ∗∗��� ���
∗�
∗���
��
��
��������� �� ��� �
������
where:�
�� � � �����
(5) ��������
��
(5) �; � � ��� �
�� � ∗�� � ����� �� � ���� � � ������ ∗ � ���…
more recently. �� � ��� � � �� ∗� ∗� ���
� �� ��� �� � �
�
… … � � ������
∗ ∗ �… �� � � � �� � � �
where:
where: �� �� � � � � … … �; �
�; �� � � � � … � � �
� � �� �� where: � …
� � � �� � � �
where:
�where:
�� � where:
���� �� ��� � � … … �� …
��� � ��
�; ���
� � � … �� � …
��� �� � �; � �; �� … � ∗��
��
�∗��
����� �
�� ���
�� �
�
� ��� …
��� �� �� …
�� ��� � ��…
���
� � � �� � � � � �
����
�� …
��
�…
�� �… ���
� ��� �� � � � �… � ��
��� � �
���� …
��… … �
� � �� � � � � …
� ��� �
����
� �����
�� �
��� …
� … �…�� …
�� �� �
where: �� �where: � … … �; � ∗ �� � �…
…�; � � � � � � …� � � � �
��where: …����� � � ∗�; � ��� … �…���∗� � ∗��� � � � ����� � � � �
��� �Since � ���� G is �a � �kxk
��� … �… ������
dimensional �� �
� �matrix
��
tor of country-specific Specifically,
foreign variables, the�the ��
� ∗ �
�� ∗,�are � constructed
plays a ��� ∗ �∗
�� as follows: ����where: ��� � � � � ��∗�� �� ���� �� ������ �
� ∑��; �� �� ���� �
��
�∗ ����
��� � ∗∗� � �
∗ �� …
�� �
� �� … � ���� �
�� ��� �� � �…� … �� � � ����
(6) ��� � � ���� � � … �a�
Specifically, Specifically, the �∗�� are constructed �� ���
� �� areas follows:
constructed as follows: By stacking By
� �stacking individual ��
individual �models:
��� �
��
Since � models:
� G is
����
a ∗kxk ����
�kxk ��
dimensional ∗� matrix
��
and in �general
� �
��
� �� willwill be ���
full rank and
��� � ��
��� �Since G�� is � a �� �� dimensional � �
�� ��matrix � and � � in
�� �� general �
� be
� ��full rank �non-
and �� no
��Since G is a kxk dimen
� ��
�� � �� �� �� ��� ���� �
urse of dimensionality when n isSince Since
relatively G G
is ais a
large,
kxk kxk thedimensional
dimensional matrix matrix and and in in general
Since
general Gwill iswill
∗Since aSince
be be
kxk G
full full
isG
dimensional
rank a is�rank
kxk a kxk
and and
dimensional non-singular.
dimensional
matrix
non-singular. and matrix matrix
in
Then Then
and
general and in by�general
in
will multiplying
general
be will
full will
rankbe byfull
be and �fullrank rankfrom
non-singular.and and thenon-singular
left,
non-singular. Then the sol TT
� ��� � �� ���Since G�multiplying
∗is (2) kxk � �� � ��� �
���∗ � ���∗ � ∑����� ∗
�� ���
�∗
������ �� ∑��� ��� ��� ∗ (2) Since ��Since �Gby� is �by �aG
multiplying ��is
kxk �aadimensional
���kxk
Since
∗ dimensional
� ��� dimensional
�G
by ∑� �by�
is� �a(2)��∑kxk �from
matrix
�� ���� matrix
�matrix
from
�
dimensional ����
the
�� and �the and
�left,and
�left,
in �thein
�in
general
general
��
matrix general
solution
the solution
will andbe
will
will
toinfull
be
the
tobethe
general
full
fullGVAR
GVAR
rank
rank
rank
willmodel
and
and
(6) and model
non-singular.
be
non-singular.
fullnon-singular.
(6)isrankobtained:
is
��obtaineand Then no
where: � � � � … �; � � � � … � � � � � � … � � � �
��� � ��� … � � � � � � … � � by multiplying by � from
by by multiplying by by � �� onfrom thethe left,thethe solution byto the multiplying GVAR ��model
by ��� is obtained:
are not estimated but specified a priori, �based by �multiplying ∗by
� from
∗�� from the theleft, left,the the solution solution �toto
��the to GVAR
the GVAR ismodel model isisobtained:
is obtained:
�� by multiplying ∗by �model from the left, the solution to the GVAR model obtained:
multiplying from left, �solution
�� �� to�the by GVAR
multiplying by is� obtained:
��� from the �
left, � the solution the GVAR model
where���� iswhere based
where�� on��trade is basedbased
shares, onon trade trade
namely, shares, shares,
the share
namely, namely,
of country
the share the j inofshare the
country total
by multiplying by of
�
�� � �� � by �multiplying
trade
j ���
in the of country
total �
by�multiplying by
�� trade
�� from � ��
i ���
andof
�� thebyfrom
country � the
left, � the
�� �� ileft,
�
��
and
from
�� the
�
solution solution
�
the ��left, to�the �the
� toGVAR
�� � the� GVAR
solution 11 �model ��to model
11 ���
the is GVAR isobtained:
obtained: obtained:
model is obtaine
the globalj economy. where: where: � � �� � … � …
�; � ∗
�; �
� � ∗…
� � � � � … � � �
� � … � � � � � … � � � � � �… � � …� � � �� �� … � � � … � �
lateral linkages incountry in the total trade of country Sincei G and is ameasured
� 11 11
kxk dimensional matrix
� �
Since
� and in G is
general
�
a �
kxk will dimensional
be � 11
full �
rank 11 and
11 matrix
11 non-singular.
�
and �
in general
Then
��� ��� 11�� �11 �� ���
∗ ∗ �11
measured in US dollar.in US dollar. ��� �� � �� �� ��� ��� � ��� ��
inmeasured
US dollar. will be full rank and non-singular.
�
Then by multiplying
hts. The choice of trade weights is based on two by relations
multiplying by ��
� from the left, the solution to the GVAR model is obtained:
In
In combiningInwith combining combining the relations with
with the the
linking relations the exogenous linking theSince linking variables
exogenous the
G is aofkxk
Since G is a
thedimensional
variables kxk by of
country-specific from
dimensional thematrix the left,
country-specific and inthe
matrix and general solution
in general will betofull will the
be rank GVAR
full rank
and non-singular.and model non-singular. is Then Then
ents an importantexogenous channel through variables which shocks of thearecountry-specific models
��
obtained:11
models to the variables
tomodels
the variablesto the in the restin
variables of in the
the the global rest of
rest model—presented
ofthethe global by
global by multiplying
multiplying
model—presented in equation
model— by � �� by
2—the from
in �equation n+1 from
the country- left, thethe
2—the left,solution
n+1 thecountry- solution to thetoGVAR the GVAR modelmodel is obtained: is obtained:
the case of ASEAN, trade linkage is much stronger models
presented in equation 2—the n+1Hence, country-specific � � ∑��� � �� ���� � � �� (7)
�� ∗ �� � �� ��
specific VECM specificmodels VECM provide models a complex provide system.
a complex system. evenHence, country-specific even country-specific �� � � are �models � � 11 � are �11 (7)
is more available in comparison to other alternative
estimated separately. estimated Aseparately. general specification A general specification for the correlation for the ��� � of � � �� �across
� shocks
correlation �of shocks different across different
e bilateral trade statistics are published annually for all �
countries/regions, nevertheless,nevertheless,
countries/regions, can be maintained. can be maintained. �� � � �� ��∗ � � �� �� � ∑��� �� ���� � � �� �� (8)
ations. Data on bilateral financial flows are either
 Measuring Fiscal and Monetary Policies Spillovers in ASEAN 35
�� � � �� ��∗ � �
�� ∗�
��������� ��
�� �����
∑ �����
� ∑������
���� �� ���� �� ��
�� (7) (7)
��� ��� � ��

Let
��� �� � � �� ���� �
�
��
� � � �� The objective of our study is to examine how
�
policy spillovers could affect the domestic economy
�� ∗�
�� � � �� ��∗ � � ��������� �� ���� ������
∑ ���
�����∑�������
� � ���� �(8)
� ��
�� of ASEAN-5 countries,
(8) (8) explicitly, spillovers effects
The on the stability of domestic price and real GDP.
The dynamics Theof thedynamics
dynamics
system are of
of thethesystem byare
system
explored explored by impulse
are explored
impulse response byresponse
analysis. analysis.
In general, the In general, the
Hence, we stimulate the shocks to foreign monetary
impulse response analysis. Into general, the mainof current and future values of each
main function ofmain function
impulse of impulse
responses responses
is to trace out theis response
trace out
of the response
current and future policy
values variable (indicating by a sudden change in
of each
function of impulse responses is to trace out the
nominal short-term interest rates) and fiscal policy
response
variable to a one-unit toofa current
variableincrease in theand
one-unit future
increase
current valuevalues
in the oneoferror
ofcurrent each
value variable
inof theoneVARerror in theItVAR
model. will bemodel. It will be
variable (indicating by a sudden change in real
to a one-unit increase in the current value of one error
estimatedour
estimated to answer to answer
researchfor our research question the impact of government
(whatofisinternal expenditure).
internal and external fiscal Specifically, we estimate
in thefor VAR model. question
It will be (what is the impact
estimated to answer forand external fiscal
GIRFs of ASEAN-5 countries’ real GDP and inflation
and monetary our andresearch
monetary
spillover on ASEAN question
spillover on (what
countries real is
ASEAN the impact
countries
GDP?). of internal
real GDP?).
by stimulating one percentage point positive shock
and external fiscal and monetary spillover on ASEAN
However, GVAR However,
modeling GVAR modelingtoo
encompasses encompasses
many countriestoo many countriesingovernment
and variables and variables inexpenditure
a complex a complex and one percentage point
countries real GDP?).
negative shock to nominal short-term interest rates.
However,
multi-country GVAR
system. modeling
Ittoisestimate
impossible to encompasses too
estimate the orthogonalized impulse responsesthe proposed
multi-country system. It is impossible the orthogonalized impulse responses Generally,
proposed transmission of fiscal and monetary
many countries and variables in a complex multi-
bywhich
Sims requires
(1980) which requires the impulse shocks into ASEAN-5
a set of countries takes[i16]: place quickly,
by Sims (1980) country system. the It is impossible
impulse responses beresponses
to to estimate
computed towith
be
thecomputed
respect towith
a setrespect
of to Commented [i16]:Commented There
There is no corresponding is no corresponding
reference entry reference entry
and the effects of shock are generally significant in
found. Pls. provide found. Pls. provide
orthogonalized impulse responses proposed by Simsimpulse response function Commented
orthogonalized orthogonalized
shocks. Hence,shocks.we willHence, estimatewegeneralized
will estimate generalized
impulse response function short-run;
(GIRF) however, (GIRF)the magnitude Commented
[U17R16]: and
Reference is sign
added of
[U17R16]:
effects
Reference is added
(1980) which requires the impulse responses to be
varied across countries. Despite the up and down
computed
which in
which was proposed was with
proposed
Koop, respect
Pesaran, and to
in Koop, a set
Pesaran,
Potter (1996)of orthogonalized
andfor
Potter (1996) models
non-linear for non-linear models and developed
and developed
fluctuation at the beginning forecast horizon, the
shocks. Hence, we will estimate generalized impulse
further
further in Pesaran in Pesaran
and Shin (1998)and forShin
vector(1998)
errorfor vector error
correcting models.correcting
The GIRFs areimpulse
models. The GIRFs
invariant responses
to are invariantmaintain
to a stable trend after 4 to
response function (GIRF) which was proposed in
12 quarters depending on cases. This indicates that
the ordering ofKoop,
the Pesaran,
the variables
ordering of thethe
and and
variables
countriesPotter
and the(1996)
in the countries
GVAR model. for
in thenon-linear
GVAR
The GIRFsmodel. The GIRFs
approach approach considers
ourconsiders
models are stable. Therefore, we will focus on
models and developed further in Pesaran and Shin
shockserrors
shocks to individual to individual errors outand integrates outthe
theother
effects of theusing analyzing
othertheshocks usingthe theresults
observed of only up to eight quarters.
(1998) forand integrates
vector error the effects of
correcting models. shocks
The observed
To comprehensively examine about how internal
distribution ofGIRFs are of
distribution
all the shocks invariant anytoorthogonalization.
all the shocks
without the ordering
without of the variables
any orthogonalization.
and external fiscal and monetary spillovers affect
and the countries in the GVAR model. The GIRFs
Empirical Results the ASEAN-5 economies as well as compare the
approach considers shocks to individual
Empirical Results errors and
effects of policy shocks from different sources,
integrates
Unit Root Tests out the effects of the other shocks using
Unit Root Tests we will analyze the responses of each ASEAN-5
the observed distribution of all the shocks without
theTo examine the
any orthogonalization.
To examine integration integration
properties of theproperties
individualof series,
the individual series,countries’
we implement we implement
the widely
realthe GDP
widely and inflation to these policy
shocks separately.
accepted
accepted standard standard
augmented augmented(ADF)
Dickey-Fuller Dickey-Fuller
tests and(ADF) testssymmetric
weighted and weighted
ADFsymmetric
(WS). TheADF (WS). The
Empirical Results Foreign fiscal spillovers.
length
length employed employed
for unit root testfor
is unit root by
selected testAIC.
is selected by AIC.
Real GDP of ASEAN-5 countries. In general,
Unit Root Tests a positive government expenditure shocks cause
To examine the integration properties of the an immediate increase in real GDP of ASEAN-5
individual series, we implement the widely accepted countries. This is illustrated by the GIRFs associated
standard augmented Dickey-Fuller
12 (ADF)
12 tests and
(of the first, fourth, and eighth quarter) plotted in
weighted symmetric ADF (WS). The length employed Figures 1 to 5. However, the country member’s
for unit root test is selected by AIC. responsiveness to fiscal shocks varies across
The estimation results of the unit root t-statistic countries and regions.
are reported in Tables 2 and 3 for country-specific and
foreign variables respectively. Overall, the results of
unit root test show that all of the country-specific and
foreign series are integration at the first order, namely
I(1).

Empirical Results
In this section, we present the empirical results
based on the dynamic analysis of the global VAR model
by estimating GIRFs.
 36

Table 2
Unit Root Test of Country-Specific Variables

CV AUS CH EURO INDIA INDO JP KR MAL PH SG TL UK USA

ADF test

y -2.89 -1.96854 -0.8699 -1.65031 -1.02126 0.222056 -1.8968 -1.6215 -0.70621 0.943982 -1.60254 -1.67476 -1.29153 -1.33474

Δy -2.89 -4.9405 -4.42419 -3.53086 -3.0947 -5.69522 -4.8631 -4.99247 -5.56546 -5.17295 -3.9168 -6.1681 -3.66682 -3.82332

Δp -2.89 -4.16912 -5.2009 -1.81324 -4.83514 -4.506 -2.7157 -2.84535 -4.76924 -5.00984 -6.71343 -5.49997 -2.77725 -3.08193

ΔΔp -2.89 -5.83123 -6.83072 -8.04425 -5.77607 -6.44143 -6.02548 -6.6653 -5.99796 -7.39644 -11.6811 -5.97398 -9.99969 -7.97232

e -2.89 -2.1958 -1.22565 -2.34976 -1.73471 -2.57766 -1.9575 -2.38868 -1.64108 -1.18515 -1.35118 -1.80297 -2.89952

Δe -2.89 -5.5512 -4.65697 -5.95374 -4.43395 -3.51336 -2.94113 -5.16678 -3.32592 -4.24327 -4.07825 -4.23504 -6.19177

r -2.89 -0.76351 -1.16606 -1.20891 -1.23401 -2.95866 -2.17938 -1.60354 -3.19935 -1.02088 -2.61055 -2.88422 -1.39499 -1.9443

Δr -2.89 -4.76521 -2.15966 -3.42022 -5.79383 -5.04625 -3.98994 -5.03276 -4.64639 -4.80574 -2.97336 -2.97284 -3.64786 -3.05176

g -2.89 -1.14921 -0.1912 -1.60938 -0.60295 -0.91434 -0.88944 -1.6344 -1.18008 0.12192 0.672662 -2.0532 -3.16935 -2.09107

Δg -2.89 -7.97491 -6.23147 -6.35313 -6.08602 -5.5864 -4.19613 -6.31317 -3.69662 -9.1756 -9.29713 -6.92412 -6.61992 -2.44129

WS test

y -2.55 1.619342 0.91249 -0.78083 0.37734 -0.03845 -1.45123 1.420566 0.980896 1.716686 -0.19615 1.494628 1.14655 0.643707

Δy -2.55 -5.0689 -4.65144 -3.75359 -2.96946 -4.48534 -4.98119 -4.87055 -5.5247 -5.40589 -4.02861 -6.41386 -3.92991 -3.98347

Δp -2.55 -4.37642 -5.41183 -0.75868 -4.94004 -4.29298 -2.93301 -3.12645 -4.29598 -5.14102 -6.41538 -5.70446 -2.62765 -3.18928

ΔΔp -2.55 -6.19431 -7.22419 -8.25462 -6.13676 -6.87414 -6.34211 -7.0719 -6.03252 -7.79292 -11.7974 -6.4318 -10.4122 -8.09733

e -2.55 -0.49269 0.091702 -0.41209 -0.78027 -0.30087 -2.20393 -1.25994 -0.93598 -0.54057 -0.75454 -0.39588 -0.87373

Δe -2.55 -5.7626 -4.85646 -6.17304 -4.68855 -3.75813 -3.05296 -5.39865 -3.55064 -4.30146 -3.92733 -4.26067 -6.42963

r -2.55 -1.09353 -0.88252 -0.96417 -1.61985 -1.01055 -2.43125 -1.60967 -3.4272 0.388881 -2.83701 -3.13175 -1.07567 -1.9187

Δr -2.55 -4.20962 -2.39352 -3.32124 -6.03403 -5.27712 -3.58276 -5.18509 -4.87412 -2.23903 -2.57912 -3.24293 -3.41374 -2.19225

g -2.55 1.169037 1.393493 1.162451 1.046944 -0.74829 0.396363 0.595395 -0.30732 0.963383 1.540598 0.510812 1.114692 0.496358

Δg -2.55 -8.22469 -6.21071 -6.5475 -6.48817 -5.26854 -4.4099 -6.5005 -3.55186 -9.30006 -9.02585 -7.00018 -6.87179 -2.61955

Source: authors’ calculation

T.M.L. Dau, et al
Table 3
Unit Root Test of Foreign Variables

CV AUS CH EURO INDIA INDO JP KR MAL PH SG TL UK USA

ADF test

y -2.89 -1.62028 -1.38208 -1.54469 -1.35558 -1.67912 -1.05818 -0.96302 -1.65861 -1.64052 -1.42384 -1.46598 -1.04692 -1.01453

Δy -2.89 -3.84374 -4.2194 -4.45657 -4.10821 -4.8216 -4.21366 -4.33496 -4.54746 -4.67641 -4.66512 -4.76433 -3.69774 -3.82272

Δp -2.89 -4.97632 -2.96133 -4.6958 -4.63525 -4.97851 -4.72209 -4.93841 -4.86252 -4.88553 -4.40154 -4.66733 -4.21093 -4.98886

ΔΔp -2.89 -6.92657 -5.61088 -6.86522 -6.76633 -7.00408 -6.82497 -6.77649 -6.13956 -6.91798 -6.30744 -6.62281 -7.56229 -6.84461

e -2.89 -1.62319 -1.37604 -2.48671 -1.23752 -1.62624 -1.58729 -1.53108 -1.67716 -1.8802 -1.88632 -1.73951 -1.63569 -0.79807

Δe -2.89 -4.21436 -3.7273 -5.4741 -3.56542 -4.04034 -4.3649 -4.11373 -3.40715 -3.0567 -3.67767 -3.15306 -3.16819 -4.06537

r -2.89 -1.14666 -1.33692 -1.14216 -1.0992 -1.35661 -1.09329 -1.10048 -1.32509 -1.41918 -1.62365 -1.16643 -1.13733 -0.91138

Δr -2.89 -2.32041 -2.59108 -3.29604 -2.43434 -2.38182 -2.38326 -2.19147 -2.16512 -2.16049 -3.80542 -3.38411 -3.32436 -3.49753

g -2.89 -0.9294 -2.0945 -1.50157 -1.04842 -0.93651 -1.02718 -0.59658 -0.62338 -1.3051 -1.09441 -0.72418 -1.46137 -0.7377
Measuring Fiscal and Monetary Policies Spillovers in ASEAN

Δg -2.89 -5.89693 -3.59146 -5.47068 -6.09605 -5.82335 -6.01933 -5.80073 -7.11121 -6.20582 -5.55562 -5.33839 -3.46541 -5.98569

WS test

y -2.55 0.474405 0.707011 0.690889 0.520344 0.487298 0.704648 0.660817 0.408657 0.581126 0.521901 0.544994 0.396117 0.709459

Δy -2.55 -4.00685 -4.37836 -4.70354 -4.25287 -5.12604 -4.41215 -4.517 -4.83869 -4.99114 -4.98472 -5.07188 -3.89628 -4.02245

Δp -2.55 -5.16589 -3.03496 -4.89605 -4.85419 -5.01337 -4.93877 -5.14413 -5.0365 -5.01891 -4.60326 -4.83757 -4.42623 -5.19644

ΔΔp -2.55 -7.33681 -5.82488 -7.2733 -7.16755 -7.36244 -7.22529 -7.17262 -6.60644 -7.31034 -6.67594 -7.01708 -7.85052 -7.24956

e -2.55 -0.32261 -0.95115 -0.13674 -0.39947 -0.61943 0.13558 -0.3036 -0.42836 -0.98849 -0.41913 -0.65 -1.12801 -0.06818

Δe -2.55 -4.38513 -3.93938 -5.6778 -3.66733 -4.18232 -4.57077 -4.29222 -3.29321 -2.94192 -3.90666 -3.12654 -3.15941 -4.27228

r -2.55 -0.81556 -1.17932 -1.0083 -0.70705 -1.39467 -0.77011 -0.65666 -1.04166 -1.30357 -0.28647 -0.609 -0.93501 -0.53759

Δr -2.55 -2.19148 -2.267 -2.2 -2.30928 -2.10596 -2.13496 -1.8054 -2.01842 -1.96937 -3.86518 -2.5633 -2.71705 -3.22198

g -2.55 1.535679 1.156413 1.303339 1.602852 1.716025 1.526358 1.477339 1.392324 1.838372 1.087804 1.428514 1.624678 1.498048

Δg -2.55 -5.88633 -3.93227 -5.49075 -5.93819 -5.68499 -6.01102 -5.67489 -6.88632 -6.0629 -5.66864 -5.18905 -3.19538 -5.96678
37

Source: authors’ calculation

 38 T.M.L. Dau, et al

For the case of Indonesia, spillover effects on
its real GDP are ambiguous. While fiscal shocks
from some countries like China, Malaysia, and
the United Kingdom are accompanied by an
instantaneous increase in its real GDP, shocks from
other countries like Korea, Thailand, and United
States cause an instantaneous decrease in its real
GDP. Furthermore, the magnitude of spillover
effects on its real GDP is relatively small compared
to the other ASEAN-5 countries’. There are only
a few countries (such as China, Korea, Malaysia,
and the United Kingdom) of which fiscal policy
has significant
associated impact
(of the first, on Indonesian
fourth, and eighth real GDP.plotted in Figures 1 to 5. However, the country
quarter)
Explicitly, shocks to China’s, Malaysia’s and
the United Kingdom’s government expenditure
member’s responsiveness to fiscal shocks varies across countries and regions.
cause a maximum increase in Indonesian real
GDP of 0.05% (UK), 0.06 % (China), and 0.09%
(Malaysia), respectively, and shock to Korea one
causes a maximum decrease of 0.006%. While
the rest of the countries, such as Australia, Euro
Area, Japan, Philippine, and Thailand, have small
and insignificant fiscal spillover to Indonesian
real GDP.
Regarding the fiscal spillovers from specific
countries to Malaysia, fiscal shocks from most countries
cause an instantaneous increase in its real GDP, except
for its own fiscal shock. Its real GDP responded
strongest to shocks from China, Singapore, Thailand,
and the United Kingdom with a maximum increase of
0.5%, 0.4%, 0.25%, and 0.35%, respectively. Fiscal
spillovers from the rest of countries have little impact.

0.001
0.0008
0.0006
0.0004
0.0002
0
-0.0002 1 4 8
-0.0004
-0.0006
-0.0008
AU CN EU IN IND JP KR ML PH SG TH UK US
Source: authors’ calculation
(Source:
Figure 1. The GIRFs authors’realcalculation)
of Indonesia’s GDP to fiscal policies shocks.

Figure 1.0.006
The response of Indonesia’s real GDP to fiscal policies shocks.
0.005
0.004
0.003
For the case
0.002of Indonesia, spillover effects on its real GDP are ambiguous. While fiscal
0.001
0
shocks from some countries like 1China, Malaysia, and4the United Kingdom 8are accompanied by
-0.001
-0.002
an instantaneous-0.003
increase in its real GDP, shocks from other countries like Korea, Thailand, and

United States cause an instantaneous

AU CN EU decrease
IN IND in JP
its real
KR GDP.
ML Furthermore,
PH SG TH the
UK magnitude
US of
Source: authors’ calculation
spillover effects on its real GDP
Figure is relatively
2. The Source:
GIRFs small compared
authors’
of Malaysia’s GDP to to
calculation
real thepolicies
fiscal othershocks.
ASEAN-5 countries’.

There are only a Figure 2. The response

few countries (such asofChina,
Malaysia’s real
Korea, GDP to fiscal
Malaysia, policies
and the Unitedshocks.
Kingdom) of

which fiscal policy has significant impact on Indonesian real GDP. Explicitly, shocks to China’s,
 -0.001
-0.002
-0.003

AU CN EU IN IND JP KR ML PH SG TH UK US
Measuring Fiscal and Monetary Policies Spillovers in ASEAN 39

Source: authors’ calculation

For the case of Philippines, our evidence shows Australia, the Euro Area, Japan, and India are relatively
that a positive fiscal
Figure 2. The response of Malaysia’sweak.
shock to most of the countries real GDP to fiscal policies shocks.
is accompanied by an instantaneous increase in Regarding to the external fiscal spillovers to
Philippine’s real GDP, except for fiscal shocks from Singapore’s real GDP, it increases instantaneously
Indonesia and Malaysia which cause the decrease in response to an external positive fiscal shock from
in Philippine’s real GDP by 0.16% and 0.14%,
Regarding the fiscal spillovers from specificmost countries, except for the ones from Australia,
countries to Malaysia, fiscal shocks from most
respectively. However, positive responses are much Euro Area, and Malaysia. Similar to Malaysia and
stronger in comparison to negative responses. The the Philippines, Singapore’s real GDP is significantly
countries cause an instantaneous increase in its real GDP, except for its own fiscal shock. Its real
strongest positive responses are found for shocks affected by fiscal spillovers from China, Korea,
from China (+0.5%), Singapore (+0.2%), Thailand Thailand, and the United Kingdom with a maximum
GDP responded strongest to shocks from China, Singapore, Thailand, and the United Kingdom
(+0.14%), the United Kingdom (+0.2%), and the U.S. increase of 1%, 0.5%, 0.3%, and 0.5%, respectively.
(+ 0.17%). The magnitude of effects is the same to Although real GDP responded negatively to response
with a maximum increase of 0.5%, 0.4%, 0.25%,
Malaysia’s and much larger compared to Indonesia’s.
and 0.35%, respectively. Fiscal spillovers from
to expansionary fiscal policies in some cases, they are
Finally, fiscal spillovers from other countries like small and insignificant.
the rest of countries have little impact.
For the case of Philippines, our evidence shows that a positive fiscal shock to most of the
0.005
countries is accompanied by an instantaneous increase in Philippine’s real GDP, except for fiscal
0.004
shocks from Indonesia
0.003 and Malaysia which cause the decrease in Philippine’s real GDP by 0.16%
0.002
and 0.14%, respectively. However, positive responses are much stronger in comparison to negative
0.001

responses. The strongest

0 positive responses are found for shocks from China (+0.5%), Singapore
1 4 8
-0.001
(+0.2%), Thailand
-0.002
(+0.14%), the United Kingdom (+0.2%), and the U.S. (+ 0.17%). The

magnitude of effects is AU
the same
CN toEUMalaysia’s
IN INDand JP
muchKRlarger
ML compared
PH SGto Indonesia’s.
TH UK USFinally,
Source: authors’ calculation
fiscal spillovers from other countries like Australia,
Source: the calculation
authors’ Euro Area, Japan, and India are relatively
Figure 3. The GIRFs of the Philippines’ real GDP to fiscal policies shocks.
weak. Figure 3. The response of the Philippines’ real GDP to fiscal policies shocks.

0.016
0.014
0.012 19
0.01
0.008
0.006
0.004
0.002
0
-0.002 1 4 8
-0.004

AU CN EU IN IND JP KR ML PH SG TH UK US
Source: authors’ calculation
Source: authors’ calculation
Figure 4. The GIRFs of Singapore’s real GDP to fiscal policies shocks.
Figure 4. The response of Singapore’s real GDP to fiscal policies shocks.
 Although real GDP responded negatively to response to expansionary fiscal policies in some cases,
40 T.M.L. Dau, et al
they are small and insignificant.

0.012
0.01
0.008
0.006
0.004
0.002
0
-0.002 1 4 8
-0.004
-0.006

AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

Source: authors’ calculation
Figure 5. The GIRFs of Thailand’s real GDP to fiscal policies shocks.
Figure 5. The GIRFs of Thailand’s real GDP to fiscal policies shocks.
For the case of Thailand, positive shocks to other the Philippines, and Thailand’s fiscal policy has a
countries’ government expenditure have a significant significant impact on Singapore’s real GDP.
positive effect on Thai real GDP, except the one from The inflation rate of ASEAN countries. The
For the case
Indonesia (-0.038%), of Thailand,
Malaysia positive
(-0.01%), and shocks
Euro to GIRFs
other countries’ government
(of the first, fourth, andexpenditure have
eighth quarter) of the
area (-0.01%). However, different to other cases, ASEAN-5 countries’ inflation rate for one percentage
Thaiareal
significant
GDP is positive effect ontoThai
less responsive fiscalreal GDP, except
spillover thepositive
point one from Indonesia
shock (-0.038%),
to government Malaysiaof all
expenditure
from other ASEAN countries. Instead, Thailand’s real countries are presented in Figures 6 to 10. Overall,
GDP(-0.01%),
respondedand Euro to
strongest area (-0.01%).
shocks However,
from other East different
positivetofiscal
otherspillovers
cases, Thai real GDP
to specific is less also
countries
Asian countries, with an increase of 1% (by shock cause an immediate fluctuation in the price level of
fromresponsive
China), 0.5% to (by
fiscal
shockspillover fromand
from Japan), other
0.5%ASEAN countries.
ASEAN-5 Instead,
countries. Thailand’s
However, real
the sign GDP
of the effects
(by shock from Korea). The U.K. is the only country are inconclusive, and the magnitudes of the reaction
responded
outside strongest
the region to shocks
which has from other
a significant impactEast
on Asianvarycountries, with an increase of 1% (by shock
across countries.
Thailand’s real GDP (+0.5%). Regarding the effects on Indonesia’s price, the
Infrom China),ASEAN-5
conclusion, 0.5% (by shock fromreal
countries’ Japan),
GDP and
are 0.5% (by shock
results fromspillovers
show that Korea). Theof theU.K. is the onlyfiscal
expansionary
relatively responsive to external fiscal shocks, and policy from specific countries causes an instantaneous
country
positive outside
spillover the region
effects which
are found forhas
mosta significant
of the impact
increaseoninThailand’s
its inflation,real GDPfor
except (+0.5%).
some cases which
cases. Particularly, we find that all ASEAN-5 countries causes a decrease in inflation; however, the inverse
In conclusion,
responded strongest to fiscalASEAN-5
shocks from countries’
other Eastreal GDP
effectareis relatively responsive
relatively small. Fiscal to external
effects fromfiscal
China are
Asian countries, especially fiscal shock from China has still dominant by causing inflation to increase by 0.4%,
strongshocks,
impactand positive
on real GDPspillover effects countries,
of all ASEAN-5 are found for then
mostfollowed
of the cases.by theParticularly, we find
U.K. (+0.2%), andthat all
Singapore
while responsiveness to Korea fiscal policy is found (+0.1%).
ASEAN-5 countries responded strongest
significant only for Singapore and Thailand and Japan to fiscal shocks
In the case from other Eastexternal
of Malaysia, Asian fiscal
countries,
spillovers
is only significant to Thailand. Other countries, except have a significant impact on its price. Fiscal policies
especially
for the United fiscal
Kingdom,shockhave
froma China has strong
very little impact.impactofoncountries
real GDPinofthe all ASEAN-5
region stillcountries,
generate while
the most
Regarding fiscal spillover between ASEAN countries, important external fiscal spillover source to Malaysia’s
responsiveness
our evidence to Korea
shows that fiscal
the fiscal policy
policy is found significant
of member inflation,only suchforasSingapore
China’s and and Japan’s
Thailandcausing
and a
countries could cause significant spillover on some maximum increase in the inflation rate of 0.3% and
other members (not all). For instance, Indonesia’s 0.2% respectively. Additionally, fiscal policies of other
21
real GDP respond significantly to fiscal shock from countries in ASEAN (Singapore: +0.2% and Thailand:
Singapore and Malaysia, while fiscal spillovers from +0.2%) also generate significant spillover to Malaysia’s
Singapore and Thailand have a strong impact on inflation.
 specific countries also cause an immediate fluctuation in the price level of ASEAN-5 countries.

However, the sign of the effects are inconclusive, and the magnitudes of the reaction vary across
Measuring Fiscal and Monetary Policies Spillovers in ASEAN 41
countries.

0.005

0.004

0.003

0.002

0.001

0
1 4 8
-0.001

-0.002

except for some cases

AU which
CN causes
EU INa decrease
IND JPin inflation;
KR ML however,
PH SG the TH
inverse
UK effect
US is relatively

Source: authors’ calculation

small. Fiscal effects from China areSource:
still dominant
authors’ by causing inflation to increase by 0.4%, then
calculation
Figure 6. The GIRFs of Indonesia’s inflation rate to fiscal policies shocks.
followed by the U.K.
Figure (+0.2%),
6. The GIRFsand Singapore (+0.1%).
of Indonesia’s inflation rate to fiscal policies shocks.

0.004
0.003
Regarding
0.002
the effects on Indonesia’s price, the results show that spillovers of the
0.001
expansionary fiscal policy from specific countries causes an instantaneous increase in its inflation,
0
-0.001 1 4 8
-0.002 22
-0.003
-0.004
-0.005

AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

Source: authors’ calculation
Figure 7. The GIRFs of Malaysia’s inflation rate to fiscal policies shocks.
Figure 7. The GIRFs of Malaysia’s inflation rate to fiscal policies shocks.

In the
The effect casespillovers
of fiscal of Malaysia, external
from other fiscal spillovers
countries have aofsignificant
the beginning impactThe
forecast horizon. on direction
its price.of
on Philippines’s inflation are diverse and remains weak. fluctuation is ambiguous, and only fiscal shocks from
TheFiscal policies
strongest of countries
positive in theonregion
effect are found shocksstill
fromgenerate
somethe most important
countries external
have a significant fiscal
effect spillover
on its inflation.
Singapore (+ 0.05%), Japan (+ 0.05%), the U.K. (+ Singapore responded the most to fiscal spillovers from
sourceand
0.05%), tothe
Malaysia’s inflation,
U.S. (+ 0.02%). such asresponse
The negative China’s and Japan’s
India causing
(increasing by a0.23%),
maximum increase
Indonesia in the
(increasing
of the Philippines’ inflation is found on shocks from by 0.2%), Malaysia (increasing by 0.2%), and Korea
Thailand (-0.12%),
inflation rate of China
0.3% (- and0.11%),
0.2% and Malaysia Additionally,
respectively. (increasing by 0.1%),
fiscal while of
policies China andcountries
other Japan cause in a
(-0.06%). decrease of 0.1%.
In Singapore’s case, foreign fiscal spillovers
ASEAN (Singapore: +0.2% and Thailand: +0.2%) also generate significant spillover to Malaysia’s
cause significant fluctuation of its inflation rate at
inflation.
 -0.001

AU CN EU IN IND JP KR ML PH SG TH UK US
42 T.M.L. Dau, et al
Source: authors’ calculation
0.0006
Figure 8. The GIRFs of The Philippines’s inflation rate to fiscal policies shocks.
0.0004
0.0002
0
-0.0002 1 4 8
The effect of fiscal spillovers from other countries on Philippines’s inflation are diverse
-0.0004
-0.0006
and remains weak. The strongest positive effect are found on shocks from Singapore (+ 0.05%),
-0.0008
-0.001
Japan (+ 0.05%), the U.K. (+ 0.05%), and the U.S. (+ 0.02%). The negative response of the
AU CN EU IN IND JP KR ML PH SG TH UK US
Philippines’ inflation is found on shocks from Thailand (-0.12%), China (- 0.11%), and Malaysia
Source: authors’ calculation
Source: authors’ calculation
Figure 8. The GIRFs of The Philippines’s inflation rate to fiscal policies shocks.
(-0.06%).
Figure 8. The GIRFs of The Philippines’s inflation rate to fiscal policies shocks.
0.005
0.004

The effect
0.003 of fiscal spillovers from other countries on Philippines’s inflation are diverse
0.002
and remains0.001
weak. The strongest positive effect are found on shocks from Singapore (+ 0.05%),
0
Japan (+ 0.05%), the U.K. (+ 0.05%), and the U.S. (+ 0.02%). The negative response of the
-0.001 1 4 8

Philippines’-0.002
inflation is found on shocks from Thailand (-0.12%), China (- 0.11%), and Malaysia

(-0.06%). AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

0.005
Source: authors’ calculation
0.004Figure 9. The GIRFs of Singapore’s inflation rate to fiscal policies shocks.

Figure 9.0.003
The GIRFs of Singapore’s inflation rate to fiscal policies shocks.
0.002
The responsiveness of Thailand’s inflation to In conclusion, spillovers of foreign expansionary
0.001 are ambiguous. The responses
external fiscal shocks fiscal policy on ASEAN-5 countries’ inflation is
on some cases fluctuate 0 intensively at the beginning ambiguous. While external and internal spillovers
of forecast horizon. The
-0.001 results show1 that the strongest cause an increase in price
4 8 in Indonesia, Malaysia, and
positive effect is observed from the U.S. (increasing by 24 Singapore, negative effects are found to be stronger in
-0.002
0.2%), United Kingdom (increasing by 0.07%), China the case of the Philippines. External spillovers (from
(increasing by 0.11%), and the Philippines (increasing outside region) cause an increase in Thailand’s inflation
AU CN EU IN IND JP KR ML PH SG TH UK US
by 0.08%). The US fiscal policy has a stronger while internal spillover (from inside the association)
impact on Thailand compared to China. Shocks from cause a decrease. Similar to the responsiveness of
Indonesia, Malaysia, and Singapore cause Source: authors’ real
a negative calculation
GDP, China’s fiscal policy still generate the
response, namely, an instantaneous fall in Thailand most important source of ASEAN-5 countries’ price
inflation byFigure
0.13%,9. 0.18%, and 0.2% respectively.
The GIRFs of Singapore’s inflation fluctuation, while
rate to Japan’s
fiscal andshocks.
policies Korea’s are only

24
 to fiscal spillovers from India (increasing by 0.23%), Indonesia (increasing by 0.2%), Malaysia

(increasing by 0.2%), and Korea (increasing by 0.1%), while China and Japan cause a decrease of
Measuring Fiscal and Monetary Policies Spillovers in ASEAN 43
0.1%.

0.0015

0.001

0.0005

0
1 4 8
-0.0005

-0.001

-0.0015

AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

Source: authors’ calculation
Figure 10. The GIRFs of Thailand’s inflation rate to fiscal policies shocks.
Figure 10. The GIRFs of Thailand’s inflation rate to fiscal policies shocks.
important for some cases. Outside the region, only Real GDP of ASEAN-5 countries. The GIRFs
fiscal spillovers from United Kingdom and the U.S. are (of the first, fourth, and eighth quarters) of ASEAN-5
also significant for some member countries. However, countries’ real GDP to internal and external monetary
The responsiveness
their magnitudes are smaller thanof Thailand’s
China’s. inflation
The rest of to external
shocks fiscalinshocks
are plotted Figures are
11 toambiguous.
15. The
the countries in our sample have little fiscal spillover For Indonesia, a negative percentage point shocks
in responses on some cases fluctuate intensively at the
ASEAN-5 countries. beginning
to other advancedofWestern
forecast horizon.short-run
economy’s The results
interest
Finally, we also found significant internal fiscal rate (Australia, Euro Area, Japan, United Kingdom,
show that
spillovers fromtheother
strongest
ASEAN-5positive effect is
countries on observed
the from
and U.S.)thecauses
U.S. (increasing by 0.2%),
an instantaneous United
fall in Indonesia
members’ real GDP and price. This finding is plausible real GDP (from -0.02% to -0.08%), while the ones
Kingdom
since ASEAN-5 (increasing by 0.07%),
economies Chinalinked,
are closely (increasingfrom
by 0.11%), and thecountries
other ASEAN+3 Philippines
cause(increasing by
an instantaneous
especially international trade between these countries. increase. Interestingly, Indonesian real GDP reacted
In0.08%).
most ASEAN-5 countries,
The US fiscal tophas
policy trading impactmost
partners
a stronger significantly
on Thailand to monetary
compared to China.shock fromfrom
Shocks other
are ASEAN countries. This indicates trade channel ASEAN countries, such as the maximum increase of
becomes an important channel in fiscal spillovers
Indonesia, Malaysia, and Singapore cause a negative 0.06% caused by
response, Thailand’s
namely, monetary shock,
an instantaneous fall then
in
into member countries. This finding suggests the followed by Singapore (+0.04%), the Philippines
importance
Thailand of policy coordination
inflation betweenand
by 0.13%, 0.18%, member (+0.04%), and Malaysia (+0.01%). However, the
0.2% respectively.
countries. magnitude of the responsiveness is relatively small
Foreign monetary spillovers. In this section, we compared to the ones of other member countries.
are going to analyze the dynamic response of ASEAN-5 The effects of foreign monetary spillovers to
countries’ real GDP and inflation to spillovers of Malaysian real GDP are ambiguous. On the one
internal and external expansionary monetary policies hand, shocks from advanced economies (Euro Area,
25
(indicated by a negative percentage point shocks to US, and Japan) and some ASEAN member countries
specific countries’ nominal short-term interest rate). (Singapore, Philippines) cause an instantaneous
Overall, the effects of monetary spillover on ASEAN-5 decrease in Malaysia’s real GDP. Externally, the cause
domestic real GDP and inflation are ambiguous. It may of maximum decreases of 0.24%, 0.16%, and 0.13% in
cause an increase or decrease in concerned variables Malaysia’s real GDP are found on shock from the Euro
depending on the cases. Additionally, the country Area, U.S., and Japan. Regarding the internal spillover
member’s responsiveness to monetary shocks varies between other ASEAN countries and Malaysia, the
across countries and regions. They generally become Philippines and Singapore monetary shocks cause a
stable after eight quarters. Hence, we will present the maximum decreases of 0.25% and 0.1% respectively.
estimated results up to eight quarters. On the other hand, positive response of Malaysia’s real
 Real GDP of ASEAN-5 countries. The GIRFs (of the first, fourth, and eighth quarters) of

ASEAN-5 countries’ real GDP to internal and external monetary shocks are plotted in Figures 11
44 T.M.L. Dau, et al
to 15.

0.0008
0.0006
0.0004
0.0002
0
-0.0002 1 4 8
-0.0004
-0.0006
-0.0008
-0.001

AU CN EU IN IND JP KR ML PH SG TH UK US
Source: authors’ calculation
Source: authors’ calculation
Figure 11. The GIRFs of Indonesia’s real GDP to monetary policies shocks.
Figure 11. The GIRFs of Indonesia’s real GDP to monetary policies shocks.
0.0025
For0.002
Indonesia, a negative percentage point shocks to other advanced Western economy’s
0.0015
short-run interest
0.001 rate (Australia, Euro Area, Japan, United Kingdom, and U.S.) causes an
0.0005
0
instantaneous
-0.0005fall in Indonesia
1 real GDP (from -0.02%
4 to -0.08%), while
8 the ones from other
-0.001
ASEAN+3 countries cause an instantaneous increase. Interestingly, Indonesian real GDP reacted
-0.0015
-0.002
-0.0025
most significantly to monetary shock from other ASEAN countries, such as the maximum increase
-0.003

of 0.06% caused by
AU Thailand’s
CN EU monetary
IN IND shock,
JP thenMLfollowed
KR PH by Singapore
SG TH UK (+0.04%),
US the
Source: authors’ calculation
Philippines (+0.04%), and Malaysia (+0.01%). However, the magnitude of the responsiveness is
Source: authors’ calculation
Figure 12. The GIRFs of Malaysia’s real GDP to monetary policies shocks.
relatively small compared
Figure to the
12. The onesofofMalaysia’s
GIRFs other member
real countries.
GDP to monetary policies shocks.

GDP to monetary shock are

The effects of also found
foreign for somespillovers
monetary other from
to the U.S.,real
Malaysian United
GDP Kingdom, Singapore,
are ambiguous. On the China,
countries, such as ASEAN countries (Indonesia and Japan, and EU, with a maximum decrease of 0.16%,
Thailand), India,shocks
one hand, and United
fromKingdom,
advancedofeconomies
which, the (Euro
0.16%,
Area,0.15%,
US, 0.13%, 0.11%,
and Japan) andsome
and 0.11%ASEAN
respectively.
strongest effects are found from Indonesia (+0.2%), In contrast to the case of the Philippines, foreign
Indiamember
(+0.13%), and Thailand (+0.07%). expansionary
countries (Singapore, Philippines) cause an instantaneous monetary
decreasespillovers to Singapore’s
in Malaysia’s real
For the Philippines, the negative effect of monetary 27 real GDP seem to generate more positive effects
spillover
GDP.seems to be much
Externally, stronger
the cause than the positive
of maximum than
decreases of negative
0.24%, ones.
0.16%, andRegarding the positivereal
0.13% in Malaysia’s effects,
ones. Its real GDP drops in response to external Singapore’s real GDP is the most responsive to
expansionary monetary
GDP are found shocksfrom
on shock from themost
EuroofArea,
the U.S.,
monetary
and Japan.shock from Indonesia
Regarding the internal(+0.6%),
spilloverIndia
countries, except for the ones from Indonesia, India, (+0.4%), Korea (+0.2%), Thailand (+0.2%), and US
Malaysia, and other
between Thailand
ASEANwhichcountries
cause anand increase in the(+0.2%).
Malaysia, Monetary
Philippines spillovers monetary
and Singapore from China, Australia,
shocks
Philippines’s real GDP by 0.13%, 0.09%, 0.05%, and EU, and the Philippines reduce Singapore’s real GDP
0.03% respectively.
cause a maximum Regarding the of
decreases negative
0.25% spillover by 0.3%, 0.2%,
and 0.1% respectively. On the 0.12%
otherand 0.15%
hand, respectively.
positive response
effects, we found the strongest responses are to shocks
of Malaysia’s real GDP to monetary shock are also found for some other countries, such as

ASEAN countries (Indonesia and Thailand), India, and United Kingdom, of which, the strongest

effects are found from Indonesia (+0.2%), India (+0.13%), and Thailand (+0.07%).
 AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

Measuring Fiscal and Monetary Policies Spillovers in ASEAN 45
Figure 13. The GIRFs of the Philippines’s real GDP to monetary policies shocks.

0.0015
0.001
For the Philippines, the negative effect of monetary spillover seems to be much stronger
0.0005

than the positive

0 ones. Its real GDP drops in response to external expansionary monetary shocks
1 4 8
-0.0005
from most -0.001
of the countries, except for the ones from Indonesia, India, Malaysia, and Thailand
-0.0015
which cause an increase in Philippines’s real GDP by 0.13%, 0.09%, 0.05%, and 0.03%
-0.002
respectively. Regarding the negative spillover effects, we found the strongest responses are to
AU CN EU IN IND JP KR ML PH SG TH UK US

shocks from
Source:the U.S.,
authors’ United Kingdom, Singapore, China, Japan, and EU, with a maximum
calculation
Source: authors’ calculation
Figure 13. The GIRFs of the Philippines’s real GDP to monetary policies shocks.
decrease of 0.16%, 0.16%, 0.15%, 0.13%, 0.11%, and 0.11% respectively.
Figure 13. The GIRFs of the Philippines’s real GDP to monetary policies shocks.
0.008

0.006
For0.004
the Philippines, the negative effect of monetary spillover seems to be much stronger
0.002
than the positive ones. Its real GDP drops in response to external expansionary monetary shocks
0
of the countries,1 except for the ones from
from most-0.002 4 8
Indonesia, India, Malaysia, and Thailand
-0.004
which cause an increase in Philippines’s real GDP by 0.13%, 0.09%, 0.05%, and 0.03%
-0.006
respectively. Regarding the negative spillover effects, we found the strongest responses are to
AU CN EU IN IND JP KR ML PH SG TH UK US

shocks from
Source:the U.S.,
authors’ United Kingdom, Singapore, China, Japan, and EU, with a maximum
calculation

Figure
decrease of 0.16%, 14. The
0.16%, GIRFs0.13%,
0.15%, of Singapore’s
0.11%,29real
andGDP to monetary
0.11% policies shocks.
respectively.

Similar to0.008
the case of other ASEAN countries,
the effect of monetary
0.006 spillovers on Thai real GDP
are mixed. Noticeably, a decrease in the Philippines’,
0.004
Euro Area’s, China’s, Australia’s, Singapore’s nominal
interest rate cause
0.002 a decrease in Thai real GDP, of
which, U.S. has the strongest impact, a decrease of
0
0.5%, then following by China (-0.43%),
1 Philippine
(-0.35%), India (-0.23%), EU (-0.23%). and United
-0.002
Kingdom (-0.2%).
-0.004 Positive effect is found from
shocks to nominal short-term interest rate of Indonesia
-0.006
(+0.28%), Malaysia (+0.32%), and Korea (+0.28%).
In conclusion, ASEAN-5 countries respond more
AU CN EU IN IND
strongly and significantly to expansionary monetary
shocks from advanced economies (i.e., to a decrease
in nominal short-term interest rate), such as Euro
Area, Japan, United Kingdom, and U.S. than from
China. However, these monetary spillover effects from
both advanced economies and China on ASEAN’s
real GDP are generally negative. Additionally, we
also found evidence of negative monetary spillovers
originating
4 from member countries
8 to ASEAN member
countries, that is, monetary spillovers from Singapore
and Philippines had caused inverse impact on another
member’s real GDP. The dominance of negative effects
of external and internal expansionary monetary policies
on ASEAN economies suggest several implications:
JP KR ML PH SG TH UK US
First, theoretically, according to Mundell-Flemming
model (1662 and 1963), expansionary monetary policy

29
46 Source: authors’ calculation T.M.L. Dau, et al

has not onlyFigure 14. impact

a positive The GIRFs
on homeofoutput
Singapore’s
but also real
forGDP
policytocoordination
monetary in policies shocks. especially
the association,
leads to home currency depreciation. This, in turn, has in monetary policymaking and exchange rate regime
two offsetting impacts on the foreign countries’ output as the monetary spillover within ASEAN countries is
(i.e., on the one hand, the increase in home country’s more significant and stronger in comparing to fiscal
In boost
output will contrast
demandto the case of
for foreign the Philippines,
countries’ output foreign expansionary monetary spillovers to
spillovers.
as well; on the other hand, home country’ currency Inflation of ASEAN countries. The GIRFs (of
Singapore’s real GDP seem to generate more positivefirst,
depreciation generates a beggar-thy-neighbor effect the fourth,
effects andnegative
than eighth quarters) of the ASEAN-5
ones. Regarding the
to foreign countries). This implies that exchange rate countries’ inflation rate for one negative percentage
channels effects,
positive are important for shock
Singapore’s realtransmission
GDP is the tomost point shock to
responsive to monetary
nominal short-term
shock frominterest rate of
Indonesia
ASEAN economies. The results are plausible since specific countries are presented in Figures 16 to 20.
most of ASEAN economies are export-based economy The monetary spillover effects on ASEAN countries
(+0.6%), India (+0.4%), Korea (+0.2%), Thailand (+0.2%),
and the fluctuation of exchange rate play a key role
and US (+0.2%). Monetary spillovers
are also mixed. Monetary spillovers could cause a rise
in their export. In short, monetary spillovers effect to or a fall in ASEAN-5 countries’ inflation depending
from
ASEAN China,
economyAustralia, EU, andrate
through exchange thechannels
Philippines
are reduce Singapore’s
on cases. Additionally,real GDP by 0.3%,
the magnitude 0.2%,
of the reaction
stronger than the demand channel. Second, negative remains limited and non-significant, and vary across
0.12% and 0.15% respectively.
spillovers within ASEAN countries suggest a need countries.

0.008

0.006

0.004

Inflation
0.002 of ASEAN countries. The GIRFs (of the first, fourth, and eighth quarters) of the

0
ASEAN-5 countries’ inflation1 rate for one negative 4percentage point shock8 to nominal short-term
-0.002

interest rate-0.004
of specific countries are presented in Figures 16 to 20. The monetary spillover effects
-0.006
on ASEAN countries are also mixed. Monetary spillovers could cause a rise or a fall in ASEAN-
AU CN EU IN IND JP KR ML PH SG TH UK US
5 countries’ inflation depending on cases. Additionally, the magnitude of the reaction remains
Source: authors’ calculation
Source: authors’ calculation
limited and non-significant,
Figure 15. and vary across
The GIRFs countries.
of Thailand’s real GDP to monetary policies shocks.

Figure 15. The GIRFs of Thailand’s real GDP to monetary policies shocks.
0.005
0.004
0.003
Similar
0.002 to the case of other ASEAN countries, the effect of monetary spillovers on Thai
0.001
real GDP are mixed.
0 Noticeably, a decrease in the Philippines’, Euro Area’s, China’s, Australia’s,
-0.001 1 4 8
Singapore’s-0.002
nominal interest rate cause a decrease in Thai real GDP, of which, U.S. has the
-0.003
strongest impact, a decrease of 0.5%, then following by China (-0.43%), Philippine (-0.35%), India
AU CN EU IN IND JP KR ML PH SG TH UK US
Source: authors’ calculation
Source: authors’ calculation
Figure 16. The GIRFs of Indonesia’s inflation rate to monetary policies shocks.
30
Figure 16. The GIRFs of Indonesia’s inflation rate to monetary policies shocks.
Measuring Fiscal and Monetary Policies Spillovers in ASEAN 47

The impact of monetary spillovers on Indonesia’s
inflation seems to be stronger than on its real GDP. It
causes a significant
0.002 instantaneous surge in domestic
inflation. Like other cases, the effects are mixed. The
0.001reactions are found for response to
strongest positive
monetary spillover0 from the U.S. (+0.31%), Singapore
(+0.23%), and India (+0.2%). A1drop in Indonesia’s
-0.001 against China (-0.26%) and the
inflation is observed
Philippine (-0.1%).
-0.002
For Malaysia, our evidence suggests that the
-0.003
negative effects of foreign expansionary spillovers are
stronger than-0.004
the positive ones. The strongest negative
response is found the shock from the Philippines
(-0.32%), U.S. (-0.3%), AU Indonesia
CN (-0.29%),
EU IN theIND
U.K. JP
(-0.2%), Japan (-0.15%), and E.U. (-0.1%). Positive
responses are found only for shocks to the nominal
interest rate of Korea (+0.12%), Singapore (+0.12%),
India (+0.08%), Thailand (+0.08%), and Australia
(+0.02%).
Compared to the case of other ASEAN countries,
the
4 responsiveness of the8 Philippines’ inflation to
monetary shocks is relatively small. It positively
responded strongest to shock from Malaysia (+0.06%),
Thailand (+0.06%), Indonesia (+0.05%), and Japan
(+0.04%). The negative reactions are found against
the U.K. (-0.07%), China (-0.05%), the U.S. (-0.04%),
and Singapore (-0.03%).
KR ML PH SG TH UK US

Source: authors’ calculation

0.002
Figure 17. The GIRFs of Malaysia’s inflation rate to monetary policies shocks.
0.001

0
For Malaysia, our evidence
1 suggests that 4the negative effects of
8 foreign expansionary
-0.001

spillovers are stronger than the positive ones. The strongest negative response is found the shock
-0.002

-0.003
from the Philippines (-0.32%), U.S. (-0.3%), Indonesia (-0.29%), the U.K. (-0.2%), Japan (-
-0.004

0.15%), and E.U. (-0.1%).

AU
Positive
CN EU
responses
IN IND
are
JP
found
KR
only
ML
for
PH
shocks
SG
toTHthe UK
nominal
US
interest rate

of Korea Source: authors’ Singapore

(+0.12%), calculation (+0.12%), India (+0.08%), Thailand (+0.08%), and Australia
Source: authors’ calculation
Figure 17. The GIRFs of Malaysia’s inflation rate to monetary policies shocks.
(+0.02%).
Figure 17. The GIRFs of Malaysia’s inflation rate to monetary policies shocks.

For 0.0006
Malaysia, our evidence suggests that the negative effects of foreign expansionary
0.0004
spillovers are stronger than the positive ones. The strongest negative response is found the shock
0.0002
0
from the Philippines (-0.32%),
1 U.S. (-0.3%), Indonesia
4 (-0.29%), the U.K.
8 (-0.2%), Japan (-
-0.0002

0.15%), and-0.0004
E.U. (-0.1%). Positive responses are found only for shocks to the nominal interest rate
-0.0006
of Korea (+0.12%),
-0.0008
Singapore (+0.12%), India (+0.08%), Thailand (+0.08%), and Australia

(+0.02%). AU CN EU IN IND JP KR ML PH SG TH UK US
Source: authors’ calculation
0.0006
Source: authors’ calculation
Figure 18. The GIRFs of the Philippines’ inflation rate to monetary policies shocks.
0.0004
Figure 18. The GIRFs of the Philippines’ inflation rate to monetary policies shocks.
0.0002
0
1 4 8
-0.0002
48 T.M.L. Dau, et al

The effects of monetary shock to specific countries India (-0.21%), the Philippines (-0.2%), the U.S.
on Singapore are still ambiguous. However, it shows (-0.15%), and Korea (-0.07%).
that the positive impact is much stronger. The shock In conclusion, our results show significant internal
from most of the countries causes an increase in and external monetary spillovers to ASEAN countries’
Singapore’s price. The strongest impact is found for price. Explicitly, negative shock to other countries’
the shock from Indonesia (+0.35%), the U.S. (+0.27%), nominal short-term interest rate cause a significant
China (+0.2%),Compared
Malaysiato(+0.2%),
the caseand
of E.U.
other(+0.16%).
ASEAN countries, the in
fluctuation responsiveness of theinflation.
ASEAN countries’ Philippines’
However,
For Thailand, monetary spillovers seem to have the effects are inconclusive. The two offsetting impacts
littleinflation
impact to
onmonetary shocks
its inflation. is relatively
Our results showsmall.
that It of
positively
internal responded
and externalstrongest
monetarytospillovers
shock from indicate
the negative impact is much stronger. Positive for a need of regional and global policy coordination.
Malaysia (+0.06%), Thailand (+0.06%), Indonesia Similar (+0.05%), to and Japan (+0.04%).
the monetary Theeffect
spillovers negativeon real
reactions are found on shocks from Singapore
(+0.11%), China (+0.11%), Indonesia GDP, advanced economies’ inflation rate is also more
reactions are found against the U.K. (+0.06%),
(-0.07%), China (-0.05%),
affected than the U.S. (-0.04%), and Singapore
China.
and Japan (+0.06%), while negative responses
are (-0.03%).
observed for cases from the U.K. (-0.23%),

0.004
0.003
0.002
0.001
0
-0.001 1 4 8
-0.002
-0.003
-0.004
-0.005

AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

Source: authors’ calculation
Figure 19. The GIRFs of Singapore’s inflation rate to monetary policies shocks.
Figure 19. The GIRFs of Singapore’s inflation rate to monetary policies shocks.
0.001
The effects of monetary shock to specific countries on Singapore are still ambiguous.
0.0005
However, it shows
0 that the positive impact is much stronger. The shock from most of the
1 4 8
countries-0.0005
causes an increase in Singapore’s price. The strongest impact is found for the shock
-0.001
from Indonesia
-0.0015 (+0.35%), the U.S. (+0.27%), China (+0.2%), Malaysia (+0.2%), and E.U.

(+0.16%).-0.002
-0.0025

AU CN EU IN IND JP KR ML PH SG TH UK US

Source: authors’ calculation

Source: authors’ calculation
Figure 20. The GIRFs of Thailand’s inflation rate to monetary policies shocks.

Figure 20. The GIRFs of Thailand’s inflation rate to monetary policies shocks.

For Thailand, monetary spillovers seem to have little impact on its inflation. Our results
Measuring Fiscal and Monetary Policies Spillovers in ASEAN
Conclusion and Policy Recommendation
By employing GVAR modeling to estimate the
GIRFs of ASEAN-5 countries’ real GDP and inflation
to expansionary fiscal and monetary shocks to countries
of interest, we can analyze the dynamic response of
these variables on the external fiscal and monetary
spillovers. The empirical estimations show several
important findings. Generally, the overall evidence
shows that both internal and external fiscal and
monetary spillovers are significant for all the cases of
ASEAN-5 countries.
Regarding fiscal spillovers, expansionary fiscal
shocks generally cause a significant increase in real
GDP of ASEAN-5 countries. However, these impacts
on ASEAN-5 countries’ inflation are ambiguous.
Additionally, our results show that fiscal spillovers
from other East Asian countries (in particular, China) to
ASEAN-5 countries are much stronger than those from
the other advanced Western countries. This indicates
that trade is an important channel in shock transmission
as trade between ASEAN countries and East Asian
countries are more tightly integrated compared to its
trade with the rest of the world.
In contrast to fiscal spillover, the effects of monetary
spillovers are inconclusive on both ASEAN countries’
real GDP and inflation. Additionally, monetary
spillover from the U.S. and European countries are
larger than those from East Asian countries.
Regarding policy spillovers within ASEAN
countries, we find sufficient evidence of internal fiscal
and monetary spillovers within ASEAN countries,
although internal monetary spillover seems to be
stronger than internal fiscal spillovers. At country level,
out of the five ASEAN countries included in our study,
Indonesia’s variables of interest are less affected by
external policy shocks. Last but not the least, there is
a negative effect of expansionary monetary spillovers
from ASEAN member countries, China, and advanced
economies on ASEAN-5 real GDP. This implies that
the exchange rate is an important channel of shock
transmission into ASEAN economies.
As mentioned, our results show significant internal
and external policy spillovers in ASEAN-5 countries’
economy. The international spillovers created by
national economic policies are not always negative,
they can be positive. For example, Korinek (2017)
found that national economic policies can create
Pareto efficiency internationally as long as certain
49
conditions are satisfied, such as national policymakers
act competitively and possess a sufficient set of policy
instruments to control these policies and there are no
imperfections in the world market. In reality, however,
it is unrealistic for such condition to be met. This
implies that in such an interconnected world where
ASEAN countries are actively participating, national
economic policies would lead to inverse international
spillover effects. Hence, there is a need for appropriate
policies to mitigate the inverse impacts of policy
spillovers.
First, regarding the policy coordination between
the ASEAN countries, as ASEAN countries are
economically linked, and evidence of internal policy
spillovers between member countries are found in
our study, ASEAN policymakers should take into
consideration the possible impacts that their policy
may have on other member countries. This emphasizes
the importance of coordination among policymakers
of all ASEAN countries in order to reduce to inverse
impacts of intra-regional policy spillovers.
Second, at the regional and global level, the
significant impacts of the regional fiscal spillover
and global financial spillovers suggest that there is
an urgent need of policy coordination at all level
(the sub-regional, regional, and global levels).
Especially, policy coordination at all levels in
monetary policymaking is crucial, as negative
monetary spillovers on ASEAN-5 real GDP are found
for shocks from many countries and regions, such
as ASEAN member countries, China, and advanced
economies. To realize these policy needs, the role
of international organizations, such as ADB and
IMF, may be important, as they could create a room
that member countries’ policymaker can work and
coordinate with each other in policy decision making.
References
Albagli, E., Ceballos, L., Claro, S., & Romero, D. (2018).
Channels of US Monetary Policy Spillovers into
International Bond Markets (BIS Working Paper No
719). Basel: Bank for International Settlements
Auerbach, A. J., & Gorodnichenko, Y. (2012). Output
spillovers from fiscal policy. In A. Alesina & F. Giavazzi
(Eds.), Fiscal policy after the financial crisis (pp. 63–98).
Chicago: University of Chicago Press.
Auerbach, A. J., & Gorodnichenko, Y. (2013). Output
spillovers from fiscal policy. American Economic
Review, 103(3), 141–146.
50
Belke, A., & Osowski, T. (2016). Measuring fiscal spillovers
in EMU and beyond: A global VAR approach (CPES
Working Document No. 428). Brussels: Centre for
European Policy Studies.
Bluedorn, J., & Bowdler, C. (2011). The open economy
consequences of U.S. monetary policy. Journal of
International Money and Finance, 30(2), 309–336.
Canova, F. (2005). The transmission of US shocks to Latin
America. Journal of Applied Econometrics, 20(2),
229–251.
Caporale, G., & Girardi, A. (2011). Fiscal spillovers in the
Euro area (DIW Discussion Paper No. 1164). Berlin:
German Institute for Economic Research.
Capennelli, C., & Houser, C. (2009). Implications of a
changing environment for monetary policy in East Asia.
Kobe University Economic Review, 55, 115-129
Chen, J., Mancini-Griffoli, T., & Sahay, R. (2014).
Spillovers from United States monetary policy on
emerging markets: Different this time? (Working
Papers WP/14/240). Washington, D.C.: International
Monetary Fund.
Chen, Q., Filardo, A., He, D., & Zhu, F. (2015). Financial
crisis, US unconventional monetary policy and
international spillovers (Working Papers WP/15/85).
Washington, D.C.: International Monetary Fund.
Dees, S., di Mauro, F., Pesaran, M., & Smith, L. (2007).
Exploring the international linkages of the Euro area: A
global VAR analysis. Journal of Applied Econometrics,
22(7), 1–38.
Dragomirescu-Gaina, C., & Philippas, D. (2015). Strategic
interactions of fiscal policies in Europe: A global
VAR perspective. Journal of International Money and
Finance, 59, 49–76.
Faust, J., & Rogers, J. (2003). Monetary policy’s role in
exchange rate behavior. Journal of Monetary Economics,
50(7), 1403–1424.
Faust, J., Rogers, J., Swanson, E., & Wright, J. (2003).
Identifying the effects of monetary policy shocks on
exchange rates using high frequency data. Journal of
the European Economic Association, 1(5), 1031–1057.
Fic, T. (2013). The spillover effects of unconventional
monetary policies in major developed countries on
developing countries (Working Paper No. 131 ST/
ESA/2013/DWP/131). New York: UN/DESA.
Fleming, J. M. (1962). Domestic financial policies
under fixed and floating exchange rates. Staff Papers
(International Monetary Fund) Vol. 9, No. 3 (Nov.,
1962), 369–380
Ganelli G., & Tawk, N. (2016). Spillovers from Japan’s
unconventional monetary policy to emerging Asia:
a global VAR approach (Working Paper No. 16/99).
Washington, D.C.: International Monetary Fund.
Georgiadis, G. (2015a). Examining asymmetries in the
transmission of monetary policy in the Euro area:
T.M.L. Dau, et al
Evidence from a mixed cross-section global VAR model.
European Economic Review, 5, 195–215.
Georgiadis, G. (2015b). Determinants of global spillovers
from US monetary policy (Working Paper Series No.
1854). Frankfurt: European Central Bank.
Goujard, A. (2013). Cross-country spillovers from fiscal
consolidations (OECD Economics Department Working
Papers No 1099 ECO/WKP(2013)91). Paris: OECD.
Ivanova, A., & Weber, S. (2011). Do fiscal spillovers
matter? (Working Paper No. 11/211). Washington, D.C.:
International Monetary Fund.
Hebous, S., & Zimmermann, T. (2013). Estimating the
effects of coordinated fiscal actions in the Euro area.
European Economic Review, 58, 110–121.
Kim, S., & Roubini, N. (2000). Exchange rate anomalies
in the industrial countries: A solution with a structural
VAR approach. Journal of Monetary Economics, 45
(2000) , 561-586.
Kim, S. (2001). International transmission of U.S. monetary
policy shocks: Evidence from VAR’s. Journal of
Monetary Economics, 48 (2001) , 339-372.
Koop, G., Pesaran, M. H., & Potter, S. M. (1996). Impulse
response analysis in nonlinear multivariate models.
Journal of Econometrics, 74, 119–147.
Korinek, A. (2017). Currency Wars or Efficient Spillovers?
A General Theory of International Policy Cooperation.
(Working Papers No 17/25). Washington, D.C.:
International Monetary Fund
Nickel, C., & Vansteenkiste, I. (2013). The international
spillover of fiscal spending on financial variables. In F.
di Mauro & H. E. Pesaran (Eds.), The GVAR handbook:
structure and applications of a macro model of the global
economy for policy analysis (pp. 182–192). Oxford:
Oxford University Press.
Nobili, A., & Neri, S. (2006). The transmission of monetary
policy shocks from the US to the Euro area (Temi di
discussion). Rome: Bank of Italy.
Miyajima, K., Mohanty, M. S., & Yetman, J. (2014).
Spillovers of US unconventional monetary policy to
Asia: The role of long-term interest rates (BIS Working
Papers No. 478). Basel: Bank for International
Settlements.
Mackowiak, B. (2007). External shocks, US monetary
policy and macroeconomic fluctuations in emerging
markets. Journal of Monetary Economics, 54(8),
2512–2520.
Mundell, R. A. (1963). Capital mobility and stabilization
policy under fixed and flexible exchange rates. Canadian
Journal of Economic and Political Science. 29 (4),
475–485.
Pesaran, M. H., &Shin, Y. (1998). Generalized impulse
response analysis in linear multivariate models.
Economics Letters, 58, 17–29.
Measuring Fiscal and Monetary Policies Spillovers in ASEAN
Pesaran, M., Schuermann, T., & Weiner, S. M. (2004).
Modeling regional interdependencies using a global
error-correcting macroeconometric model. Journal of
Business and Economic Statistics, 22(2), 129–162.
OECD. (2009). The effectiveness and scope of fiscal
stimulus. In OECD Economic Outlook, Interim
Report March 2009 (pp. 105–150). Paris: OECD
Publishing.
51
Ricci-Risquete, A., & Ramajo-Hernández, J. (2015).
Macroeconomic effects of fiscal policy in the European
Union: A GVAR model, Empirical Economics, 48,
1587–1617.
Sims C. (1980). Macroeconomics and Reality, Econometrica
48, 1-48.
Taylor, J. B. (1993). Macroeconomic policy in a world
economy: From econometric design to practical
operation. New York: W.W. Norton.